Collaborative Computing Networking

Applications and Worksharing 13th

International Conference
CollaborateCom 2017 Edinburgh UK
December 11 13 2017 Proceedings Imed
Romdhani
Visit to download the full and correct content document:
https://textbookfull.com/product/collaborative-computing-networking-applications-and-
worksharing-13th-international-conference-collaboratecom-2017-edinburgh-uk-decem
ber-11-13-2017-proceedings-imed-romdhani/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Collaborative Computing Networking Applications and

Worksharing 11th International Conference
CollaborateCom 2015 Wuhan November 10 11 2015 China
Proceedings 1st Edition Song Guo
https://textbookfull.com/product/collaborative-computing-
networking-applications-and-worksharing-11th-international-
conference-collaboratecom-2015-wuhan-november-10-11-2015-china-
proceedings-1st-edition-song-guo/

Medical Image Understanding and Analysis 21st Annual

Conference MIUA 2017 Edinburgh UK July 11 13 2017
Proceedings 1st Edition María Valdés Hernández

https://textbookfull.com/product/medical-image-understanding-and-
analysis-21st-annual-conference-miua-2017-edinburgh-uk-
july-11-13-2017-proceedings-1st-edition-maria-valdes-hernandez/

Intelligent Human Computer Interaction 9th

International Conference IHCI 2017 Evry France December
11 13 2017 Proceedings 1st Edition Patrick Horain

https://textbookfull.com/product/intelligent-human-computer-
interaction-9th-international-conference-ihci-2017-evry-france-
december-11-13-2017-proceedings-1st-edition-patrick-horain/

Distributed Computing and Internet Technology 13th

International Conference ICDCIT 2017 Bhubaneswar India
January 13 16 2017 Proceedings 1st Edition Padmanabhan
Krishnan
https://textbookfull.com/product/distributed-computing-and-
internet-technology-13th-international-conference-
icdcit-2017-bhubaneswar-india-january-13-16-2017-proceedings-1st-
Intelligent Computing Theories and Application 13th
International Conference ICIC 2017 Liverpool UK August
7 10 2017 Proceedings Part I 1st Edition De-Shuang
Huang
https://textbookfull.com/product/intelligent-computing-theories-
and-application-13th-international-conference-
icic-2017-liverpool-uk-august-7-10-2017-proceedings-part-i-1st-
edition-de-shuang-huang/

Service Oriented Computing 15th International

Conference ICSOC 2017 Malaga Spain November 13 16 2017
Proceedings 1st Edition Michael Maximilien

https://textbookfull.com/product/service-oriented-computing-15th-
international-conference-icsoc-2017-malaga-spain-
november-13-16-2017-proceedings-1st-edition-michael-maximilien/

Smart Computing and Communication Second International

Conference SmartCom 2017 Shenzhen China December 10 12
2017 Proceedings 1st Edition Meikang Qiu

https://textbookfull.com/product/smart-computing-and-
communication-second-international-conference-
smartcom-2017-shenzhen-china-december-10-12-2017-proceedings-1st-
edition-meikang-qiu/

Advances in Computer Entertainment Technology 14th

International Conference ACE 2017 London UK December 14
16 2017 Proceedings 1st Edition Adrian David Cheok

https://textbookfull.com/product/advances-in-computer-
entertainment-technology-14th-international-conference-
ace-2017-london-uk-december-14-16-2017-proceedings-1st-edition-
adrian-david-cheok/

Shell structures : theory and applications. Volume 4.

Proceedings of the 11th International Conference "Shell
Structures: Theory and Applications" (SSTA 2017),
October 11-13, 2017, Gdansk, Poland 1st Edition
Wojciech Pietraszkiewicz
https://textbookfull.com/product/shell-structures-theory-and-
applications-volume-4-proceedings-of-the-11th-international-
conference-shell-structures-theory-and-applications-
Imed Romdhani
Lei Shu
Hara Takahiro
Zhangbing Zhou
Timothy Gordon
Deze Zeng (Eds.)
252

Collaborative Computing:
Networking, Applications
and Worksharing
13th International Conference, CollaborateCom 2017
Edinburgh, UK, December 11–13, 2017
Proceedings

123
Lecture Notes of the Institute
for Computer Sciences, Social Informatics
and Telecommunications Engineering 252

Editorial Board
Ozgur Akan
Middle East Technical University, Ankara, Turkey
Paolo Bellavista
University of Bologna, Bologna, Italy
Jiannong Cao
Hong Kong Polytechnic University, Hong Kong, Hong Kong
Geoffrey Coulson
Lancaster University, Lancaster, UK
Falko Dressler
University of Erlangen, Erlangen, Germany
Domenico Ferrari
Università Cattolica Piacenza, Piacenza, Italy
Mario Gerla
UCLA, Los Angeles, USA
Hisashi Kobayashi
Princeton University, Princeton, USA
Sergio Palazzo
University of Catania, Catania, Italy
Sartaj Sahni
University of Florida, Florida, USA
Xuemin Sherman Shen
University of Waterloo, Waterloo, Canada
Mircea Stan
University of Virginia, Charlottesville, USA
Jia Xiaohua
City University of Hong Kong, Kowloon, Hong Kong
Albert Y. Zomaya
University of Sydney, Sydney, Australia
More information about this series at http://www.springer.com/series/8197
Imed Romdhani Lei Shu
•

Hara Takahiro Zhangbing Zhou

•

Timothy Gordon Deze Zeng (Eds.)

•

Collaborative Computing:
Networking, Applications
and Worksharing
13th International Conference, CollaborateCom 2017
Edinburgh, UK, December 11–13, 2017
Proceedings

123
Editors
Imed Romdhani Zhangbing Zhou
Edinburgh Napier University China University of Geosciences
Edinburgh Beijing
UK China
Lei Shu Timothy Gordon
Guangdong University of Petrochemical University of Nebraska–Lincoln
Technology Lincoln
Maoming UK
China
Deze Zeng
Hara Takahiro School of Computer Science
Osaka University China University of Geosciences
Osaka Wuhan, Hubei
Japan China

ISSN 1867-8211 ISSN 1867-822X (electronic)

Lecture Notes of the Institute for Computer Sciences, Social Informatics
and Telecommunications Engineering
ISBN 978-3-030-00915-1 ISBN 978-3-030-00916-8 (eBook)
https://doi.org/10.1007/978-3-030-00916-8

Library of Congress Control Number: 2018954768

© ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2018
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are
believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors
give a warranty, express or implied, with respect to the material contained herein or for any errors or
omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
 Preface

We are delighted to introduce the proceedings of the 13th edition of the 2017 European
Alliance for Innovation (EAI) International Conference on Collaborative Computing:
Networking, Applications and Worksharing (CollaborateCom). This conference has
brought together researchers, developers, and practitioners from around the world who
are leveraging and developing collaboration between distributed teams of humans,
computer applications, and/or autonomous robots to achieve higher productivity and
produce joint products that would have been impossible to develop without the con-
tributions of multiple collaborators.
The technical program of CollaborateCom 2017 consisted of 65 papers, in oral
presentation sessions during the main conference tracks. The three-day conference
included ten sessions. Aside from the high-quality technical paper presentations, the
technical program also featured one keynote speech titled “Converged Technology for
Intelligent Airport Services.” The keynote speaker was Prof. Henry Wang from the
University of West London, UK.
Coordination between the Steering Committee and the Organizing Committee was
essential for the success of the conference. We sincerely appreciate the support and
guidance of the general chairs: Prof. Takahiro Hara, Lei Shu and Imed Romdhani. We
would like to thank the program chairs: Zhangbing Zhou, Timothy Gordon, and
Takahiro Hara, who supervised the review process of the technical papers and com-
piled a high-quality technical program. We greatly appreciate the excellent support and
hard work of the local chair, Prof. Imed Romdhani. We are also grateful to the con-
ference manager, Dominika Belisova, for her support, and all the authors who sub-
mitted their papers to the CollaborateCom 2017 conference.
We strongly believe that the CollaborateCom conference provides a good forum for
all researchers, developers, and practitioners to discuss all science and technology
aspects that are relevant to collaborative computing. We also expect that the future
CollaborateCom conference will be as successful and stimulating as indicated by the
contributions presented in this volume.

August 2018 Imed Romdhani

Lei Shu
Hara Takahiro
Zhangbing Zhou
Timothy Gordon
Deze Zeng
 Organization

Steering Committee
Imrich Chlamtac EAI/Create-Net, Italy
Song Guo University of Aizu, Japan
Bo Li The Hong Kong University of Science and Technology,
SAR China
Xiaofei Liao Huazhong University of Science and Technology, China

Organizing Committee
Honorable Chair
Qinghua Zhang Guangdong University of Petrochemical Technology,
(President) China

General Co-chairs
Lei Shu Guangdong University of Petrochemical Technology,
China/University of Lincoln, UK
Takahiro Hara Osaka University, Japan
Imed Romdhani Edinburgh Napier University, UK

TPC Co-chairs
Timothy Gordon University of Lincoln, UK
Zhangbing Zhou China University of Geosciences, Beijing,
China/TELECOM SudParis, France

General Track Co-chairs

Liehuang Zhu Beijing Institute of Technology, China
Jianwei Niu Beihang University, China

Track 1 Co-chairs
Jianwu Wang University of Maryland, Baltimore County, USA
Zhuofeng Zhao North China University of Technology, China

Track 2 Co-chairs
Yucong Duan Hainan University, China
Antonella Longo University of Salento, Italy
VIII Organization

Track 3 Co-chairs
Shigang Yue University of Lincoln, UK
Farshad Arvin University of Manchester, UK

Track 4 Co-chairs
Helge Janicke De Montfort University, UK
Leandros Maglaras De Montfort University, UK/ABMS Open University,
Switzerland

Local Co-chairs
Jiafu Wan South China University of Technology, Guangzhou, China
Jianfeng Huang Guangdong University of Petrochemical Technology,
China

Workshops Co-chairs
Peng Li University of Aizu, Japan
Chunsheng Zhu The University of British Columbia, Canada

Publicity Co-chairs
Oli Mival Edinburgh Napier University, UK
Weipeng Jing Northeast Forestry University, China
Daichi Amagata Osaka University, Japan

Posters and PhD Track Chair

Mithun Mukherjee Guangdong University of Petrochemical Technology,
China

Publication Co-chairs
Fei Lei Beijing University of Technology, China
Deze Zeng China University of Geosciences, Wuhan, China

Demo Chair
Amjad Mehmood Guangdong University of Petrochemical Technology,
China

Poster/Demo Co-chair
Sanjay Madria Missouri University of Science and Technology, USA

Sponsorship Co-chairs
Li Yan Guangdong University of Petrochemical Technology,
China
Guangjie Han Hohai University, China
 Organization IX

Web Chair
Zhiqiang Huo University of Lincoln, UK

Special Sessions
Special Session: Internet of Things
Special Session Chair
Ahmed Al-Dubai Edinburgh Napier University, UK

Special Session: Collaborative Next Generation Networking

Special Session Co-chairs
Deze Zeng China University of Geosciences, Wuhan, China
Heng Qi Dalian University of Technology, China

Special Session: Collaboration Techniques in Data-Intensive Cloud Computing

Special Session Co-chairs
Shaojie Qiao Chengdu University of Information Technology, China
Cheqing Jin East China Normal University, China
Hongzhi Wang Harbin Institute of Technology, China

Special Session: Collaborative Manufacturing System and Machinery Fault

Diagnosis
Special Session Co-chairs
Ailin Yu Guangdong University of Technology, China
Chilou Zhou South China University of Technology, China
Jianfeng Huang Guangdong University of Petrochemical Technology,
China

Workshop: Vehicular Telematics over Heterogeneous Wireless Networks

Workshop Chair
Daxin Tian Beihang University, China

Conference Manager
Dominika Belisova European Alliance for Innovation
X Organization

Technical Program Committee

Ali Emre Turgut Middle East Technical University, Turkey
Abdelrahman Osman University of Tabuk, Saudi Arabia
Elfaki
Alex Norta Tallinn University of Technology, Estonia
Antonella Longo University of Salento, Italy
Anna Lisa Ferrara University of Surrey, UK
Athanasios Giannetsos University of Surrey, UK
Aggeliki Tsohou Ionian University, Greece
Antonis Tzounis University of Thessaly, Greece
Artemios Voyiatzis TU Wien, Austria
Bin Li Guangzhou Maritime University, China
Christos Xenakis University of Piraues, Greece
Chao Tong Beihang University, China
Dayang Sun The University of British Columbia, Canada
Francois Siewe De Montfort University, UK
Fekade Getahun Addis Ababa University, Ethiopia
Guolin He South China University of Technology, China
Guangquan Xu Tianjin University, China
Georgios V. Lioudakis National Technological University of Athens, Greece
Georgios Karopoulos National and Kapodistrian University of Athens, Greece
Honghao Gao Shanghai University, China
Helge Janicke De Montfort University, UK
Haoming Li Freescale Semiconductors, Inc., USA
Haipeng Dai Nanjing University, China
Iryna Yevseyeva De Montfort University, UK
Jianlong Xu Shantou University, China
Jianqiang Wang Tsinghua University, China
Jian Wang Inria, France
Jin Liu Wuhan University, China
Jin Li Guangzhou University, China
Joe Tekli Lebanese American University (LAU), Lebanon
Joel Rodrigues University of Beira Interior, Portugal
Kutluk Bilge Arikan Atilim University, Turkey
Mengxuan Li China Ship Development and Design Center, China
Maria Karyda University of the Aegean, Greece
Marco Comerio University of Milano-Bicocca, Italy
Mohamed Amine Ferrag Guelma University, Algeria
Meikang Qiu Pace University, USA
Mingliang Xu Zhengzhou University, China
Man Hon Cheung The Chinese University of Hong Kong, SAR China
Nanjangud C. Narendra MS Ramaiah University of Applied Sciences, India
Nianjun Joe Zhou IBM T. J. Watson Research Center, USA
Nikos Tziritas Chinese Academy of Sciences, China
Ouyang Zhenchao Beihang University, China
 Organization XI

Payam Zahadat University of Graz, Austria

Panagiotis Andriotis University of the West of England, UK
Pavlina Fragkou Technological Institute of Athens, Greece
Qiang Duan Pennsylvania State University, USA
Shuichang Liu Hunan University of Technology, China
Shangguan Wei Beijing Jiaotong University, China
Shangguang Wang Beijing University of Posts and Telecommunications,
China
Shizhan Chen Tianjin University, China
Stephan University of Leicester, UK
Reiff-Marganiec
Simon Watson University of Manchester, UK
Shasha Mo Beihang University, China
Sotiris Moschoyiannis University of Surrey, UK
Song Guo Hong Kong Polytechnic University, SAR China
Tomas Krajnik Czech Technical University, Czech Republic
Tiago Cruz University of Coimbra, Portugal
Vasilis Sourlas University College London, UK
Wentao Wang Guangzhou City Polytechnic, China
Wencan Zhang Foshan University, China
Weiwei Fang Beijing Jiaotong University, China
Yingchun Long Shaoguan University, China
Yingjun Chen Guangdong University of Petrochemical Technology,
China
Yucong Duan Hainan University, China
Yunchuan Sun Beijing Normal University, China
Ying He De Montfort University, UK
Ying Li Beihang University, China
Yu Shui Deakin University, Australia
Liangyong Qi Qufu Normal University, China
Limin Sun Chinese Academy of Sciences, China
Long Cheng Virginia Tech, USA
Lu Liu University of Derby, UK
Le Dong University of Electronic Science and Technology of China
(UESTC), China
Xiujuan Liang Guangdong Ocean University, China
Xiaolei Ma University of Washington, USA
Xiaobing Sun Yangzhou University, China
Xiong Li Hunan University of Science and Technology, China
Xinkai Wu Cal Poly Pomona, USA
Zhipeng Cai Georgia State University, USA
Zheli Liu Nankai University, China
 Contents

Main Track

Sentiment Analysis of Chinese Words Using Word Embedding

and Sentiment Morpheme Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Jianwei Niu, Mingsheng Sun, and Shasha Mo

Collaborative and Green Resource Allocation in 5G HetNet

with Renewable Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Yi Liu, Yue Gao, Shengli Xie, and Yan Zhang

On Cost Efficient Dataflow Computing Program Deployment in SDN

Managed Distributed Computing Environment . . . . . . . . . . . . . . . . . . . . . . 27
Yating Zhang, Yuepeng Li, Long Zheng, and Deze Zeng

A Short Review of Constructing Noise Map Using

Crowdsensing Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Xuhui Zhang, Lei Shu, Zhiqiang Huo, Mithun Mukherjee, and Yu Zhang

Research of Positioning Tracking on Dynamic Target Based on the Integral

Complementing Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Jianrong Xu, Guiping Lu, Jingsheng Lin, and Jianfeng Huang

LWTP: An Improved Automatic Image Annotation Method Based

on Image Segmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Jianwei Niu, Shijie Li, Shasha Mo, and Jun Ma

A Comparative Analysis of Content Delivery Capability for Collaborative

Dual-Architecture Network. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Xuan Liu, Peng Yang, Yongqiang Dong, and Syed Hassan Ahmed

AXE: Objects Search in Mobile Volunteered Service. . . . . . . . . . . . . . . . . . 73

Yao Wu, Wenjuan Liang, Yuncheng Wu, Hong Chen, and Cuiping Li

Modeling the Impacts of WiFi Signals on Energy Consumption

of Smartphones. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Yuxia Sun, Junxian Chen, and Yong Tang

Constrained Route Planning Based on the Regular Expression . . . . . . . . . . . 98

Jing Wang, Huiping Liu, and Zhao Zhang

A Human-Machine Collaborative Detection Model for Identifying

Web Attacks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Yong Hu, Bo Li, Weijing Ye, and Guiqin Yuan
XIV Contents

A Reinforcement Learning Based Workflow Application Scheduling

Approach in Dynamic Cloud Environment . . . . . . . . . . . . . . . . . . . . . . . . . 120
Yi Wei, Daniel Kudenko, Shijun Liu, Li Pan, Lei Wu, and Xiangxu Meng

p-Faster R-CNN Algorithm for Food Detection. . . . . . . . . . . . . . . . . . . . . . 132

Yanchen Wan, Yu Liu, Yuan Li, and Puhong Zhang

PUED: A Social Spammer Detection Method Based on PU Learning

and Ensemble Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
Yuqi Song, Min Gao, Junliang Yu, Wentao Li, Lulan Yu, and Xinyu Xiao

Position vs. Attitude: How Topological Factors Influence Our Difference

in the Attitudes on Online Interrelationships? A Case Study
with Language Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Bo Wang, Yingjun Sun, Yuexian Hou, Dawei Song, and Ruifang He

Route-Oriented Participants Recruitment in Collaborative Crowdsensing . . . . 164

Shu Yang, Jinglin Li, Quan Yuan, and Zhihan Liu

Management Node Selection Based on Cloud Model

in a Distributed Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
Hanyi Tang, Qibo Sun, and Jinglin Li

CNTE: A Node Centrality-Based Network Trust Evaluation Method . . . . . . . 186

Xiang Yuan, Qibo Sun, and Jinglin Li

Minimum Cost Load Migration in Distributed Software Defined Networks. . . 197

Kuangyu Qin, Chuanhe Huang, N. Ganesan, and Kewei Liu

Hybrid-Aware Collaborative Multipath Communications

for Heterogeneous Vehicular Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
Yana Liu, Wei Quan, Jinjie Zeng, Gang Liu, and Hongke Zhang

Internet of Things (IoT) and Collaboration

RETaIL: A Machine Learning-Based Item-Level Localization System

in Retail Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Xiaoyi Xu, Xiaoming Chen, Jiang Ji, Feng Chen, and Addicam V. Sanjay

A Point of Interest Recommendation Approach by Fusing Geographical

and Reputation Influence on Location Based Social Networks . . . . . . . . . . . 232
Jun Zeng, Feng Li, Junhao Wen, and Wei Zhou

An Electricity Power Collection Data Oriented Missing Data

Imputation Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Jiangqi Chen, Han Li, Ting Zhao, and He Liu
 Contents XV

Curve-Registration-Based Feature Extraction for Predictive Maintenance

of Industrial Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
Shouli Zhang, Xiaohong Li, Jianwu Wang, and Shen Su

Predicting Next Points of Interests Based on a Markov Model . . . . . . . . . . . 264

Jie Xu, Chunxiao Xing, and Yong Zhang

A Topology-Aware Reliable Broadcast Scheme for Multidimensional

VANET Scenarios. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Fengrui Liu, Chuanhe Huang, and Xiying Fan

MAC-ILoc: Multiple Antennas Cooperation Based Indoor Localization

Using Cylindrical Antenna Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
Wu Jie, Zhu Minghua, and Xiao Bo

Easing Traffic Congestion: An Improved Clustering Method for Sharing

Bike Station Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
Jian Kang, Weipeng Jing, and Chengfang Zhao

Sensor Data Classification for the Indication of Lameness in Sheep. . . . . . . . 309

Zainab Al-Rubaye, Ali Al-Sherbaz, Wanda McCormick, and Scott Turner

Collaborative Data and Workflow Management

Learning Planning and Recommendation Based on an Adaptive

Architecture on Data Graph, Information Graph and Knowledge Graph . . . . . 323
Lixu Shao, Yucong Duan, Zhangbing Zhou, Quan Zou,
and Honghao Gao

A 3D Registration Method Based on Indoor Positioning

Through Networking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
Haibo Li, Huahu Xu, Honghao Gao, Minjie Bian, and Huaikou Miao

A Load Balancing Method Based on Node Features

in a Heterogeneous Hadoop Cluster. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
Pengcheng Yang, Honghao Gao, Huahu Xu, Minjie Bian,
and Danqi Chu

Towards Collaborative Typed Resources Manipulation in Health-Care

Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
Lixu Shao, Yucong Duan, Zhangbing Zhou, Antonella Longo,
Donghai Zhu, and Honghao Gao

Constructing Search as a Service Towards Non-deterministic and Not

Validated Resource Environment with a Positive-Negative Strategy . . . . . . . . 365
Yucong Duan, Lixu Shao, Xiaobing Sun, Lizhen Cui, Donghai Zhu,
and Zhengyang Song
XVI Contents

A Passenger Flow Analysis Method Through Ride Behaviors

on Massive Smart Card Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
Weilong Ding, Zhuofeng Zhao, Han Li, Yaqi Cao, and Yang Xu

Ranking the Influence of Micro-blog Users Based on Activation

Forwarding Relationship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
Yiwei Yang, Wenbin Yao, and Dongbin Wang

UR Rank: Micro-blog User Influence Ranking Algorithm Based

on User Relationship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
Wenbin Yao, Yiwei Yang, and Dongbin Wang

Using the MapReduce Approach for the Spatio-Temporal Data Analytics

in Road Traffic Crowdsensing Application . . . . . . . . . . . . . . . . . . . . . . . . . 405
Sandhya Armoogum and Shevam Munchetty-Chendriah

An Efficient Critical Incident Propagation Model for Social Networks

Based on Trust Factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416
XiaoMing Li, Limengzi Yuan, ChaoChao Liu, Wei Yu, Xue Chen,
and Guangquan Xu

Impact of the Important Users on Social Recommendation System . . . . . . . . 425

Zehua Zhao, Min Gao, Junliang Yu, Yuqi Song, Xinyi Wang,
and Min Zhang

Analysis on Communication Cost and Team Performance in Team

Formation Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
Weijun Chen, Jing Yang, and Yang Yu

An Efficient Black-Box Vulnerability Scanning Method

for Web Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 444
Haoxia Jin, Ming Xu, Xue Yang, Ting Wu, Ning Zheng, and Tao Yang

Collaboration with Artificial Intelligence

Collaborative Shilling Detection Bridging Factorization

and User Embedding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
Tong Dou, Junliang Yu, Qingyu Xiong, Min Gao, Yuqi Song,
and Qianqi Fang

Integrating User Embedding and Collaborative Filtering for Social

Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 470
Junliang Yu, Min Gao, Yuqi Song, Qianqi Fang, Wenge Rong,
and Qingyu Xiong

CORB-SLAM: A Collaborative Visual SLAM System for Multiple Robots . . . 480

Fu Li, Shaowu Yang, Xiaodong Yi, and Xuejun Yang
 Contents XVII

Security and Trustworthy

An Efficient and Privacy Preserving CP-ABE Scheme

for Internet-Based Collaboration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493
Jinmiao Wang and Bo Lang

An Efficient Identity-Based Privacy-Preserving Authentication Scheme

for VANETs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 508
Jie Cui, Wenyu Xu, Kewei Sha, and Hong Zhong

ProNet: Toward Payload-Driven Protocol Fingerprinting via Convolutions

and Embeddings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 519
Yafei Sang, Yongzheng Zhang, and Chengwei Peng

Formal Verification of Authorization Policies for Enterprise Social

Networks Using PlusCal-2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 530
Sabina Akhtar, Ehtesham Zahoor, and Olivier Perrin

High Performance Regular Expression Matching on FPGA . . . . . . . . . . . . . 541

Jiajia Yang, Lei Jiang, Xu Bai, and Qiong Dai

Adaptive Carving Method for Live FLV Streaming . . . . . . . . . . . . . . . . . . . 554

Haidong Ge, Ning Zheng, Lin Cai, Ming Xu, Tong Qiao, Tao Yang,
Jinkai Sun, and Sudeng Hu

Android App Classification and Permission Usage Risk Assessment . . . . . . . 567

Yidong Shen, Ming Xu, Ning Zheng, Jian Xu, Wenjing Xia, Yiming Wu,
Tong Qiao, and Tao Yang

A Privacy Settings Prediction Model for Textual Posts on Social Networks . . . 578
Lijun Chen, Ming Xu, Xue Yang, Ning Zheng, Yiming Wu, Jian Xu,
Tong Qiao, and Hongbin Liu

Collaboration Techniques in Data-intensive Cloud Computing

HSAStore: A Hierarchical Storage Architecture for Computing Systems

Containing Large-Scale Intermediate Data . . . . . . . . . . . . . . . . . . . . . . . . . 591
Zhoujie Zhang, Limin Xiao, Shubin Su, Li Ruan, Bing Wei, Nan Zhou,
Xi Liu, Haitao Wang, and Zipeng Wei

Early Notification and Dynamic Routing: An Improved SDN-Based

Optimization Mechanism for VM Migration . . . . . . . . . . . . . . . . . . . . . . . . 602
Xuanlong Qin, Dagang Li, Ching-Hsuan Chen, and Nen-Fu Huang

Cloud Data Assured Deletion Based on Information Hiding

and Secondary Encryption with Chaos Sequence. . . . . . . . . . . . . . . . . . . . . 613
Yijie Chen and Wenbin Yao
XVIII Contents

Performance Analysis of Storm in a Real-World Big Data Stream

Computing Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 624
Hongbin Yan, Dawei Sun, Shang Gao, and Zhangbing Zhou

Collaborative Next Generation Networking

Shortest Path Discovery in Consideration of Obstacle in Mobile Social

Network Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 637
Dawei Sun, Wentian Qu, Shang Gao, and Li Liu

Minimize Residual Energy of the 3-D Underwater Sensor Networks

with Non-uniform Node Distribution to Prolong the Network Lifetime . . . . . 647
Gaotao Shi, Jia Zeng, Chunfeng Liu, and Keqiu Li

Combinational Meta-paths Mining for Correlation Relationship Evaluation

in Bibliographic Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 660
Qinchen Wu, Peng Wu, and Li Pan

Detecting Spammer Communities Using Network Structural Features . . . . . . 670

Wen Zhou, Meng Liu, and Yajun Zhang

Citation Based Collaborative Summarization of Scientific Publications

by a New Sentence Similarity Measure . . . . . . . . . . . . . . . . . . . . . . . . . . . 680
Chengzhe Yuan, Dingding Li, Jia Zhu, Yong Tang, Shahbaz Wasti,
Chaobo He, Hai Liu, and Ronghua Lin

Collaborative Manufacturing System and Machinery Fault Diagnosis

Exploiting User Activities for Answer Ranking in Q&A Forums. . . . . . . . . . 693

Chenyang Zhao, Liutong Xu, and Hai Huang

A Two-Level Classifier Model for Sentiment Analysis. . . . . . . . . . . . . . . . . 704

Haidong Hao, Li Ruan, Limin Xiao, Shubin Su, Feng Yuan,
Haitao Wang, and Jianbin Liu

An Effective Method for Self-driving Car Navigation based on Lidar . . . . . . 718

Meng Liu, Yu Liu, Jianwei Niu, Yu Du, and Yanchen Wan

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 729

Main Track
 Sentiment Analysis of Chinese Words
Using Word Embedding and Sentiment
Morpheme Matching

Jianwei Niu(B) , Mingsheng Sun, and Shasha Mo

State Key Laboratory of Virtual Reality Technology and Systems,

School of Computer Science and Engineering, Beihang University,
Beijing 100191, China
niujianwei@buaa.edu.cn

Abstract. Sentiment analysis has become significantly important with

the increasing demand of Natural Language Processing (NLP). A novel
Chinese Sentiment Words Polarity (CSWP) analyzing method, which
is based on sentiment morpheme matching method and word embed-
ding method, is proposed in this paper. In the CSWP, the sentiment
morpheme matching method is creatively combined with existing word
embedding method, it not only successfully retained the advantages of
flexibility and timeliness of the unsupervised methods, but also improved
the performance of the original word embedding method. Firstly, the
CSWP uses word embedding method to calculate the polarity score
for candidate sentiment words, then the sentiment morpheme match-
ing method is applied to make further analysis for the polarity of words.
Finally, to deal with the low recognition ratio in the sentiment mor-
pheme matching method, a synonym expanding step is added into the
morpheme matching method, which can significantly improve the recog-
nition ratio of the sentiment morpheme matching method. The perfor-
mance of CSWP is evaluated through extensive experiments on 20000
users’ comments. Experimental results show that the proposed CSWP
method has achieved a desirable performance when compared with other
two baseline methods.

Keywords: Sentiment polarity analysis · Word formation rule

Sentiment morpheme matching · Word embedding
Synonym expanding

1 Introduction
Recent years has witnessed the rapid development of the Internet. With the
explosive growth in the online comment number, the value of massive informa-
tion has emerged gradually. Mining the valuable information from comments
often relies on extracting the sentimental polarity of texts. In much previous
work, the sentiment analysis is done at document level. However, the sentiment
c ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2018
I. Romdhani et al. (Eds.): CollaborateCom 2017, LNICST 252, pp. 3–12, 2018.
https://doi.org/10.1007/978-3-030-00916-8_1
4 J. Niu et al.

analyzing of comments often requires sentence-level or even word-level sentiment

analysis. What’s more, the word-level sentiment analysis is the basics of all other
sentiment analyzing works.
One promising sentiment analysis method is to combine the manual tagged
corpora with the supervised learning method. In this sentiment analysis method,
a supervised learning model will be trained with tagged dataset, and then use
the trained model to execute analyzing work. For training the machine learning
models, [1] manually collected and tagged 177 negative tweets and 182 positive
tweets and, [2] manually tagged 900 sentences and 2519 words. As can be seen
that before the analysis work, lots of corpora must be manually collected and
tagged for training the supervised machine learning models. However, collecting
and tagging the massive corpora is time costly and boring. Additionally, the
training set in [1] is collected in period time, and the training set in [2] only
content limited sentiences and words. The limited size will lead these datasets
cannot include enough sentiment expression form, thereby leading the bad adap-
tive capacity in different situations.
The other promising sentiment analysis method is to use massive untagged
corpora to train unsupervised method. A typical practice of these methods is
to use the statistical methods on massive corpora to calculate the probability
of a word’s sentiment polarity. [3] exploited the PMI to calculate the polar-
ity score for the candidate sentiment words. However, the PMI is sensitive to
frequency. It means that two low frequency words can also have a high PMI
value. Additionally, [4] utilized word embedding method to convert words into
word vectors and use word vectors to calculate the similarity between candidate
words and sentiment seed sets. Word2vec tool can generate the real vectors for
words according to the context of them. However, two words which have opposite
sentiment polarity also can have similar context in sometimes.
In this paper, A novel Chinese Sentiment Words Polarity (CSWP) analyzing
method is proposed. The design of the CSWP is to exploit the word embedding
method and the sentiment morpheme matching method. The CSWP retains
the advantages about flexibility and timeliness of the unsupervised methods.
These features can lead to better practicality in real life. The performance of
CSWP is evaluated through extensive experiments on 20000 online comments.
The experimental results showed that the proposed CSWP method achieved a
desirable performance compared with other methods.
The rest paper is organized as follows. Section 2 reviews related research
about word-level sentiment analysis. Section 3 overviews the architecture of
CSWP and describes the implement of CSWP in detail. Section 4 reports a
series of experiments to evaluate the performance of CSWP, and the results are
shown in the end of this section. Finally, conclusions are driven in Sect. 5.

2 Related Work
This section will review the related work in the field of word-level sentiment
analysis. Up to now, there are two kinds of methods have been proposed for
 CSWP 5

sentiment words analysis: the supervised based method and the unsupervised
based method.
The methods based on supervised method always use some trained mathe-
matical models and some manual tagged corpora to predict the sentiment polar-
ity of words. For example, [5] considered the polarity judging question as a binary
classification problem, and used SVM models to analyze the sentiment polarity.
Moreover, [6,7] used conjunctions as the training features of machine learning
models to judge the sentiment polarity of sentiment words. In addition, morphol-
ogy information and syntactic information also have been exploited in sentiment
polarity analysis [8]. What’s more, with the huge growth in the use of microblogs
[9,10], [11,12] implemented SVM model to extract the emoticons and as these
emoticons as the judging gists for analyzing the sentiment polarity of text. They
considered that emoticons can accurately reflect the real emotion of microblogs.
As for methods based on the unsupervised method, they usually utilize mas-
sive untagged corpora and statistical methods. [13] directly used the similarity
of words to judge their sentiment polarity. [14] also used the value of similarity,
but the authors considered the similarity as the features of clustering method.
Additionally, the statistical variables such as Point Mutual Information, Coin-
cident Entropy and PageRank also have been used in sentiment analysis [3,15].
These statistical variables are exploited to calculate the sentiment polarity score
for candidate sentiment words. Because the unsupervised methods are based on
untagged corpora and statistical methods, they generally have relative low pre-
cision ratio. However, the advance of do not need manual tagged corpora leads
these methods always have strong timeliness and can adapt variable situations.

3 Methodology
This section will describe the detailed flow of the CSWP and the rationale behind
the design of the CSWP. Firstly, the overview of the CSWP will be introduced,
and then the detail of the CSWP will be described step by step. Finally, the
pseudo code will be shown in the end of this section.

3.1 System Overview

The architecture of the proposed CSWP method is shown in Fig. 1. The CSWP
method mainly contains three phases: the preprocessing phase, the initial judge
phase and the further judge phase. In the preprocessing phase, a large collection
of online comments is segmented into independent Chinese words. The segment
step is implemented by using the HanLP Chinese segment tool1 —a useful Chi-
nese text processing tool. After finished the segment step, all the adjectives are
picked out to build up candidate sentiment words set. In the initial judge phase,
the candidate sentiment words are given. The initial judge phase is responsible
to calculate polarity score for every candidate sentiment words. In the end, can-
didate words are sent into further judge phase. The further judge phase exploits
1
https://github.com/hankcs/HanLP.
6 J. Niu et al.

sentiment morpheme matching method to make further judge for the polarity of
the candidate sentiment words, and the final judging results will be output by
this phase.

CSWP

Online Candidate Word2Vec Words Similarity Word Embedding

Preprocessing
Comments Words Vectors Judging

Pre-processing Initial Judge

Sentiment Morpheme Analysis

Morphemes Matching judging Results

Further Judge

Fig. 1. Architecture of CSWP method

3.2 Preprocessing Phase

Unlike English, there are not separators between Chinese words, so the word
segmentation is a necessary step in Chinese natural language processing. The
preprocessing phase is responsible to segment the dataset and pick out all the
candidate sentiment words by using the HanLP Chinese segment tool. During
segmenting the dataset into independent words, the HanLP also tags the part-of-
speech for every candidate sentiment words. The HanLP Chinese segment tool
utilize “ad” to tag the adjectives. Then in the step of picking out the candidate
sentiment words, the preprocessing phase picks out all the adjectives as the
candidate words. The examples of Chinese words segmentation results are shown
in Table 1.

Table 1. The example of word segmentation results

3.3 Initial Judge Phase

The initial judge phase is used to calculate the polarity score for every candidate
sentiment words. The specific process of this phase can be divided into the
following two steps:

– Step 1. Generate Word Vectors

Firstly, all the words are sorted according to their frequency in the dataset,
and then twenty obvious positive candidate words and twenty obvious nega-
tive candidate words are picked out to build up positive negative seed sets.
These two seed sets will be used in step 2 to assist with calculating the polarity
 CSWP 7

score. Next, all the candidate sentiment words are converted into 200 dimen-
sional vectors by exploiting word2vec tool. The word2vec tool can convert the
words into continuous space vectors and while retain the context features for
the words.
– Step 2. Calculate the Polarity Score
Firstly, the value of similarity between candidate sentiment words and senti-
ment seed sets (including the positive seed set and the negative seed set) is
calculated. Then the similarity between candidate word and the positive seed
set is defined as the word’s positive similarity, and the similarity between a
candidate word and negative seed set is defined as the word’s negative seed
set. After getting the similarity, the polarity score of a word is calculated
by utilizing the word’s positive similarity to minus it’s negative similarity.
The positive and negative of the score implies the positive and negative of
the sentiment polarity of the word. The formula of computing similarity and
polarity score are shown in 1 and 2, respectively.
m
 k
l=1 Ail ×Bjl
sim(Ai |B) =   (1)
k 2 k 2
j=1 l=1 (Ail ) × l=1 (Bjl ) ×m
score(Ai ) = sim(Ai |C) − sim(Ai |D) (2)

where A donates the set of candidate words, and B donates the set of
sentiment seeds (positive seeds or negative seeds). Ai (i = 1, 2, . . . , n) and
Bj (j = 1, 2, . . . , m) donate the vector of the word. Vector Ai is composed
by Ai1 , Ai2 , . . . , Aik , and vector Bj is composed by Bj1 , Bj2 , . . . , Bjk . The
subscript k is the dimension of the word vector. What’s more, C donates the
positive sentiment seed set, and D donates the negative sentiment seed set.
Cj and Dj donate the vector of word, they have same structure as Bj .

3.4 Further Judge Phase

Although word embedding method can judge the polarity of sentiment words
independently, the frequency of words and the context of words can influence
the analyzing results. Moreover, the size and the content of text also will influence
the results. All above factors may lead to a reduction in performance. To mitigate
the influence of these issues, a sentiment word formation rule matching method
is proposed to assist with the word embedding method. Through utilizing the
formation rule matching method, the CSWP can correct some errors caused by
previous factors. But if you want to use formation rules to judge the polarity of
sentiment words independently, you may also get a bad result, because formation
rules always hard to be summarized, and it is impossible to collect all rules.
Therefore, an independent rule matching method cannot adapt complex real
situations.
The sentiment word formation rules which the CSWP exploits is based on
sentiment morpheme matching method and synonym expanding method. Exist-
ing work mentioned that some sentiment words express their sentiment through
8 J. Niu et al.

their interior morphemes, so exploiting some sentiment morpheme matching

rules can also analyze the sentiment polarity of Chinese words. For example,
only contains the positive morpheme , so this word
will be judged as positive sentiment word. But contains the pri-
vative morpheme and positive , furthermore, the privative
morpheme has more front position than positive morpheme in ,
so will be judged as negative sentiment word. The specific
steps of this phase can be divided into following 3 steps:

– Step1. Sentiment Morpheme Matching

After using cosine similarity to calculate the polarity score, CSWP scans every
candidate word for checking whether it contains positive morpheme, negative
morpheme and privative morpheme. In course of scanning morphemes, CSWP
will utilize the morpheme matching rules to judge the sentiment polarity of
the candidate words. If a candidate word cannot match any rule, then CSWP
will jump to step 2. If all the candidate words have finished the sentiment mor-
pheme matching, then CSWP will jump to step 3. The sentiment morpheme
matching problem can be described as follows: given a candidate sentiment
Chinese word W = w1 w2 . . . wn , where wi is the morpheme in the Chinese
word W , and subscript i is the position indicator of wi . The goal of the
method is to determine the position of wi in W . Before giving the specific
rules description, we have several conventions about symbol expression, they
are shown as follows:
1. if wi is a positive morpheme, then use pi to replace it.
2. if wi is a negative morpheme, then use ei to replace it.
3. if wi is a privative morpheme, then use ri to replace it.
So, the specific rules can be described as follows:
– Rule1: given a W , if pi ∈ W , ej ∈ / W , rk ∈
/ W ; then score = 1.
– Rule2: given a W , if pi ∈ W , ej ∈ / W , rk ∈ W and i < k; then score = 0.
– Rule3: given a W , if pi ∈ W , ej ∈ / W , rk ∈ W and k < i; then score =
−1.
– Rule4: given a W , if pi ∈ / W , ej ∈ W , rk ∈
/ W ; then score = −1.
– Rule5: given a W , if pi ∈ / W , ej ∈ W , rk ∈ W and k < j; then score = 1.
– Rule6: given a W , if pi ∈ / W , ej ∈ W , rk ∈ W and j < k; then score = 0.
– Rule7: given a W , if pi ∈ W , ej ∈ W , rk ∈ / W and j < i; then score =
−1.
– Rule8: given a W , if pi ∈ W , ej ∈ W , rk ∈ W and k < j < i; then
score = 1.
– Step 2. Synonyms Expanding
If a candidate word can’t match any rule, then the process will jump to
this step. The candidate word sent to this step will be expanded to some
synonyms by searching synonym dictionary. Then, the CSWP will use this
word’s synonyms continue to execute step 1. And if one of the synonyms can
match a rule, the polarity score of this candidate word also will be changed.
– Step 3. Output Final Results
Finally, after step 1 has finished the sentiment morpheme matching process,
 CSWP 9

the final results have come into being. The candidate words which have pos-
itive score will be marked as positive sentiment words, and the candidate
words which have negative score will be marked as negative sentiment words.
Moreover, the words which have zero score will be given up.

3.5 Pseudo Code of CSWP

Algorithm 1 explains how CSWP works. Line 2–7 corresponds the word embed-
ding judging phase. Line 8–14 corresponds the sentiment morpheme matching
phase.

Algorithm 1. Sentiment polarity judging

Input: SDP dataset, sentiment morphemes, synonym dictionary.
Output: positive sentiment word set, negative sentiment word set.
1: Let A donates the candidate word set, C donates the positive seed set, D donate
the negative seed set.
2: convert all Ai into word vectors.
3: build up C and D /*C and D are picked from A*/.
4: for all Ai do
5: use equation(1) to calculate the sim(Ai |C) and sim(Ai |D)
6: score(Ai ) = sim(Ai |C) − sim(Ai |D)
7: end for
8: for all Ai do
9: if Ai can match a rule of morpheme matching then
10: update score(Ai ) according to the rule.
11: else
12: utilize synonym expanding to expand synonyms and jump back to Line 9.
13: end if
14: end for

4 Experiments and Results

In this section, we firstly introduce the preparing of the experimental dataset,
and then we introduced a series of comparative experiments. The purpose of
these experiments is to evaluate whether the CSWP method has a desirable per-
formance in word-level Chinese sentiment analysis compared with two baseline
method.

4.1 Dataset
To validate the proposed method, 20000 online comments were extracted from
Star Data Platform2 , a text data supply platform. For the sake of convenience,
the collection of these comments is named as SDP dataset. The source of SDP
2
https://www.istarshine.com/index.php/Data/dataSurvey#platform.
10 J. Niu et al.

dataset including microblogs, news and post bar. Then, through summarizing
from the NTUSD Chinese sentiment words dictionary3 , three sentiment mor-
pheme sets were build up: a positive sentiment morpheme set, a negative senti-
ment morpheme set and a privative morpheme set. As for the synonym dictionary
utilized in CSWP, the HIT-CIR synonym dictionary4 was downloaded, this dic-
tionary was developed by HIT Center for Information Retrieval Lab. The sample
of sentiment morphemes are shown in Table 2.

Table 2. The sample of the Sentiment Morphemes

4.2 Experimental Results

In this subsection, the performance of the proposed CSWP system is evaluated

with SDP dataset. To illustrate the advantages of CSWP, we also implement
and evaluate an unsupervised baseline method [4] and a supervised method [2]
for comparison.
For illustrating the performance of the CSWP method. The 10-fold cross-
validation is performed to the three methods. In each time of experiments, 2000
comments of SDP dataset are randomly selected as the training data and the rest
are the testing data. The training dataset is used to train supervised baseline
method, and the testing dataset is used to evaluate the performance of three
methods. The final comparative results are divided into positive sentiment words
extracting results and negative sentiment words extracting results. The results
are shown in Figs. 2 and 3.

Fig. 2. Positive results Fig. 3. Negative results

3
http://www.datatang.com/data/44317.
4
http://www.ltp-cloud.com/download/.
 CSWP 11

As can be seen from Figs. 2 and 3, the CSWP achieves the best F1 value,
which is average 19.075% higher than unsupervised baseline and average 9.475%
higher than supervised baseline. Additionally, we can see through the comparison
of the first method and the CSWP that the synonym expanding step has average
improved the accuracy of CSWP by 5.53% and average improved the recall
ratio by 11.195%. It has brought significant improvement to the CSWP. The
reason of the improvement is that the step of synonym expanding can improve
the recognition ratio of the sentiment morpheme matching method. Therefore,
there are more wrong results can be corrected by this phase, and thereby the
performance of the CSWP is improved.

5 Conclusion
This paper proposed a novel Chinese Sentiment Words Polarity (CSWP) analyz-
ing method that exploits the word embedding method and the sentiment mor-
pheme matching method. Compared with the existing word embedding method,
the performance of the CSWP method has achieved the significant improvement
in analyzing the sentiment polarity of Chinese words. To illustrate the perfor-
mance of the CSWP, two baseline methods were implemented, and then the three
methods were evaluated on SDP dataset. The experimental results show that
the CSWP method leads to a desirable performance compared with the unsu-
pervised baseline method which only based on word embedding method, and it
also can be seen from results that the performance of the CSWP also better
than the supervised baseline method which based on maximum entropy model.
The advantages of the CSWP method are that it only needs little manpower in
preparing work, and it retains the advantages of flexibility and timeliness which
belong to unsupervised method, so it will be more competent than supervised
baseline method in different harsh work situation.

Acknowledgment. This work was supported by the National Natural Science Foun-
dation of China (61572060, U1536107, 61472024), and CERNET Innovation Project
(NGII20151104, NGII20160316).

References
1. Go, A., Bhayani, R., Huang, L.: Twitter sentiment classification using distant
supervision. CS224N Project Report, Stanford, 1(2009):12 (2009)
2. Fei, X., Wang, H., Zhu, J.: Sentiment word identification using the maximum
entropy model. In: 2010 International Conference on Natural Language Processing
and Knowledge Engineering (NLP-KE), pp. 1–4. IEEE (2010)
3. Turney, P.D.: Thumbs up or thumbs down?: semantic orientation applied to unsu-
pervised classification of reviews. In: Proceedings of the 40th Annual Meeting on
Association For Computational Linguistics, pp. 417–424. Association for Compu-
tational Linguistics (2002)
4. Fan, X., Li, X., Du, F., Li, X., Wei, M.: Apply word vectors for sentiment analysis
of app reviews. In: 2016 3rd International Conference on Systems and Informatics
(ICSAI), pp. 1062–1066. IEEE (2016)
12 J. Niu et al.

5. Pang, B., Lee, L., Vaithyanathan, S.: Thumbs up?: sentiment classification using
machine learning techniques. In: Proceedings of the ACL-2002 Conference on
Empirical Methods in Natural Language Processing, vol. 10, pp. 79–86. Associ-
ation for Computational Linguistics (2002)
6. Hatzivassiloglou, V., McKeown, K.R.: Predicting the semantic orientation of adjec-
tives. In: Proceedings of the Eighth Conference on European Chapter of the Asso-
ciation for Computational Linguistics, pp. 174–181. Association for Computational
Linguistics (1997)
7. Kanayama, H., Nasukawa, T.: Fully automatic lexicon expansion for domain-
oriented sentiment analysis. In: Proceedings of the 2006 Conference on Empirical
Methods in Natural Language Processing, pp. 355–363. Association for Computa-
tional Linguistics (2006)
8. Ku, L.-W., Liang, Y.-T., Chen, H.-H.: Opinion extraction, summarization and
tracking in news and blog corpora. In: Proceedings of AAAI, pp. 100–107 (2006)
9. Jansen, B.J., Zhang, M., Sobel, K., Chowdury, A.: Twitter power: Tweets as elec-
tronic word of mouth. J. Assoc. Inf. Sci. Technol. 60(11), 2169–2188 (2009)
10. Pak, A., Paroubek, P.: Twitter as a corpus for sentiment analysis and opinion
mining. In: LREc, vol. 10 (2010)
11. Kouloumpis, E., Wilson, T., Moore, J.D.: Twitter sentiment analysis: the good the
bad and the omg! Icwsm, 11(538-541), 164 (2011)
12. Huang, S., Han, W., Que, X., Wang, W.: Polarity identification of sentiment words
based on emoticons. In: 2013 9th International Conference on Computational Intel-
ligence and Security (CIS), pp. 134–138. IEEE (2013)
13. Gauch, S., Wang, J.: Corpus analysis for TREC 5 query expansion. In: TREC
(1996)
14. Wiebe, J.: Learning subjective adjectives from corpora. AAAI/IAAI 20 (2000)
15. Geng, H.T., Cai, Q.S., Kun, Y., Zhao, P.: A kind of automatic text keyphrase
extraction method based on word co-occurrence. J. Nanjing Univ. 42(2), 156–162
(2006)
 Collaborative and Green Resource
Allocation in 5G HetNet
with Renewable Energy

Yi Liu1,2(B) , Yue Gao3 , Shengli Xie1,2 , and Yan Zhang4

1
School of Automation, Guangdong University of Technology, Guangzhou, China
yi.liu@gdut.edu.cn
2
Guangdong Province Key Laboratory of IoT Information Technology,
Guangzhou, China
3
Queen Mary University of London, London, UK
yue.gao@qmul.ac.uk
4
University of Oslo, Oslo, Norway
yanzhang@ieee.org

Abstract. By deploying dense renewable-connected small-cells, hetero-

geneous networks (HetNets) are able to provide spectral and energy effi-
ciencies for 5G system. However, the small-cell base stations (BSs) may
suffer the intra-cell interferences and variabilities of renewable energy.
In this paper, we firstly introduce a collaborative architecture to deal
with intra-cell interferences and renewable uncertainty for different-tier
of users in HetNets. A stochastic optimization problem is formulated
to maximize the energy efficiency of collaborative HetNet. To solve the
problem, a centralized resources allocation algorithm is proposed based
on random variabilities of spectrum and renewable resources. Finally, the
extensive numerical results are provided to verify the effectiveness of the
proposed collaborative resources allocation method.

Keywords: Collaborative HetNet · Resource allocation

Renewable uncertainty

1 Introduction
Heterogeneous network (HetNet) is treated as a promising solution to support
tremendous number of diverse terminals and wireless services in 5G system. How-
ever, the dense deployment of small-cell BSs and terminals in HetNet will lead to
two important issues: spectrum efficiency and energy efficiency. To address these
two issues, some technologies have been intensively studied. Massive multiple-
input multiple-output (MIMO) technology have been proven to its potential of
significantly improving the spectral efficiency about 10–20 times in the same
frequency bandwidth [1]. Cognitive radio (CR) technology has been proposed to
effectively utilize the spectrum [2,3]. The CR users/devices are allowed to oppor-
tunistically operate in the frequency bands originally allocated to the primary
c ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2018
I. Romdhani et al. (Eds.): CollaborateCom 2017, LNICST 252, pp. 13–26, 2018.
https://doi.org/10.1007/978-3-030-00916-8_2
Another random document with
no related content on Scribd:
honor. The offence which provoked this assault is not even hinted at,
though it may have arisen from the troubled state of public affairs.
Captain Praa was a man of influence and dignity in the community,
an exiled Huguenot, of remarkable skill in horsemanship and arms.
In spite of all this, it appears probable that the sentiment of the
community was in sympathy with the two turbulent assaulters and
batterers, for they were fined only six shillings and three pounds
respectively. They threw themselves on the mercy of the Court, and
certainly were treated with mercy.
There are, however, few women-criminals named in the old Dutch
and early English records, and these few were not prosecuted for
any very great crimes or viciousness; the chief number were brought
up for defamation of character and slander, though men-slanderers
were more plentiful than women. The close intimacy, the ideal
neighborliness of the Dutch communities of New York made the
settlers deeply abhor all violations of the law of social kindness. To
preserve this state of amity, they believed with Chaucer “the first
vertue is to restraine and kepen wel thine tonge.”
The magistrates knew how vast a flame might be kindled by a
petty spark; and therefore promptly quenched the odious slander in
its beginning; petty quarrels were adjusted by arbitration ere they
grew to great breaches. As sung the chorus of Batavian women in
Van der Vondel’s great poem:—

“If e’er dispute or discord dared intrude,

’Twas soon by wisdom’s voice subdued.”

In spite, however, of all wariness and watchfulness and patience,

the inevitable fretfulness engendered in petty natures by a narrow
and confined life showed in neighborhood disputes and suits for
defamation of character, few of them of great seriousness and most
of them easily adjusted by the phlegmatic and somewhat dictatorial
Dutch magistrates. In a community so given to nicknaming it seems
strange to find such extreme touchiness about being called names.
Suits for defamation were frequent, through opprobrious name-
calling, and on very slight though irritating grounds. It would certainly
seem a rather disproportionate amount of trouble to bring a lawsuit
simply because you were called a “black pudding,” or a verklickker,
or tale-bearer, or even a “Turk;” though, of course, no one would
stand being called a “horned beast” or a “hay thief.” Nor was “Thou
swine” an offensive term too petty to be passed over in silence. The
terrible epithets, spitter-baard and “Dutch dough-face,” seem to
make a climax of opprobriousness; but the word moff was worse, for
it was the despised term applied in Holland to the Germans, and it
led to a quarrel with knives.
I wish to note in passing that though the Dutch called each other
these disagreeable and even degrading names, they did not swear
at each other. Profanity was seldom punished in New Amsterdam,
for practically it did not exist, as was remarked by travellers.
Chaplain Wolley told of “the usual oath” of one Dutch colonist,—the
word “sacrament.”
The colonists were impatient of insulting actions as well as words.
Sampson said in “Romeo and Juliet,” “I will bite my thumb at them,
which is a disgrace to them if they bear it;” so “finger-sticking” was a
disgrace in colonial times if unresented, and it was actionable in the
courts. The man or woman who pointed the finger of scorn at a
neighbor was pretty sure to have the finger of the law pointed at him.
The curious practice of the Dutch settlers alluded to—the giving of
nicknames—may be partly explained by the fact that in some cases
the persons named had no surname, and the nickname was really a
distinguishing name. These nicknames appear not only in the
records of criminal cases, but in official documents such as the
patents for towns, transfers of estates, civil contracts, etc. In Albany,
in 1655 and 1657, we find Jan the Jester, Huybert the Rogue,
Jacobus or Cobus the Looper, squint-eyed Harmen, the wicked
Domine. On Long Island were John the Swede, Hans the Boor,
Tunis the Fisher. In Harlem was Jan Archer the Koop-all (or buy-all).
In New York, in English days, in 1691, we find Long Mary, Old Bush,
Top-knot Betty, Scarebouch. These names conveyed no offence,
and seem to have been universally adopted and responded to.
 It would appear to a casual observer glancing over the court-
records of those early years of New York life under Dutch
supremacy, that the greater number of the cases brought before the
magistrates were these slander and libel cases. We could believe
that no other court-room ever rang with such petty personal suits; to
use Tennyson’s words, “it bubbled o’er with gossip and scandal and
spite.” But in truth slander was severely punished in all the colonies,
in New England, Virginia, Pennsylvania; and it is not to the detriment
of the citizens of New Netherland that they were more sharp in the
punishment of such offences, for it is well known, as Swift says, that
the worthiest people are those most injured by slander.
The slander cases of colonial times seem most trivial and even
absurd when seen through the mist of years. They could scarce
reach the dignity of Piers Plowman’s definition of slanders:—

“To bakbyten, and to bosten, and to bere fals witnesse

To scornie and to scolde, sclaundres to make.”

To show their character, let me give those recorded in which

Thomas Applegate of Gravesend, Long Island, took an accused part.
In 1650, he was brought up before the Gravesend court for saying of
a fellow-towns-man that “he thought if his debts were paid he would
have little left.” For this incautious but not very heinous speech he
paid a fine of forty guilders. The next year we find him prosecuted for
saying of a neighbor that “he had not half a wife.” Though he at first
denied this speech, he was ordered “to make publick
acknowledgement of error; to stand at the publick post with a paper
on his breast mentioning the reason, that he is a notorious,
scandalous person.” This brought him to his senses, and he
confessed his guilt, desired the slandered “half a wife” to “pass it by
and remit it, which she freely did and he gave her thanks.” Next
Mistress Applegate was brought up for saying that a neighbor’s wife
milked the Applegate cows. She escaped punishment by proving
that Penelope Prince told her so. As a climax, Thomas Applegate
said to a friend that he believed that the Governor took bribes. The
schout in his decision on this grave offence said Applegate “did
deserve to have his tongue bored through with a hot iron;” but this
fierce punishment was not awarded him, nor was he banished.
When the tailor of New Amsterdam said disrespectful words of the
Governor, his sentence was that he “stand before the Governor’s
door with uncovered head, after the ringing of the bell, and to declare
that he falsely and scandalously issued such words and then to ask
God’s pardon.”
The magistrates were very touchy of their dignity. Poor Widow
Piertje Jans had her house sold on an execution; and, exasperated
by the proceeding, and apparently also at the price obtained, she
said bitterly to the officers, “Ye despoilers, ye bloodsuckers, ye have
not sold but given away my house.” Instead of treating these as the
heated words of a disappointed and unhappy woman, the officers
promptly ran tattling to the Stadt Huys and whiningly complained to
the Court that her words were “a sting which could not be endured.”
Piertje was in turn called shameful; her words were termed “foul,
villanous, injurious, nay, infamous words,” and also called a
blasphemy, insult, affront, and reproach. She was accused of
insulting, defaming, affronting, and reproaching the Court, and that
she was in the highest degree reprimanded, particularly corrected,
and severely punished; and after being forbidden to indulge in any
more such blasphemies, she was released,—“bethumped with
words,” as Shakespeare said,—doubtless well scared at the
enormity of her offence, as well as at the enormity of the magistrate’s
phraseology.
The notary Walewyn van der Veen was frequently in trouble,
usually for contempt of court. And I doubt not “the little bench of
justices” was sometimes rather trying in its ways to a notary who
knew anything about law. On one occasion, when a case relating to
a bill of exchange had been decided against him, Van der Veen
spoke of their High Mightinesses the magistrates as “simpletons and
blockheads.” This was the scathing sentence of his punishment:—
“That Walewyn Van der Veen, for his committed insult, shall
here beg forgiveness, with uncovered head, of God, Justice,
 and the Worshipful Court, and moreover pay as a fine 190
guilders.”
This fine must have consumed all his fees for many a weary
month thereafter, if we can judge by the meagre lawyers’ bills which
have come down to us.
Another time the contumacious Van der Veen called the Secretary
a rascal. Thereat, the latter, much aggrieved, demanded “honorable
and profitable reparation” for the insult. The schout judged this
epithet to be a slander and an affront to the Secretary, which
“affected his honor, being tender,” and the honor of the Court as well,
since it was to a member of the Court, and he demanded that the
notary should pay a fine of fifty guilders as an example to other
slanderers, “who for trifles have constantly in their mouths curses
and abuses of other honorable people.”
Another well-known notary and practitioner and pleader in the
busy little Court held in the Stadt Huys was Solomon La Chair. His
manuscript volume of nearly three hundred pages, containing
detailed accounts of all the business he transacted in Manhattan, is
now in the County Clerk’s Office in New York, and proves valuable
material for the historiographer. He had much business, for he could
speak and write both English and Dutch; and he was a faithful,
painstaking, intelligent worker. He not only conducted lawsuits for
others, but he seems to have been in constant legal hot water
himself on his own account. He was sued for drinking and not paying
for a can of sugared wine; and also for a half-aam of costly French
wine; and he was sued for the balance of payment for a house he
had purchased; he pleaded for more time, and with the ingenuous
guilelessness peculiar to the law said in explanation that he had had
the money gathered at one time for payment, but it had somehow
dropped through his fingers. “The Court condemned to pay at
once,”—not being taken in by any such simplicity as that. He had to
pay a fine of twelve guilders for affronting both fire inspector and
court messenger. He first insulted the brandt-meester who came to
inspect his chimney, and was fined, then he called the bode who
came to collect the fine “a little cock booted and spurred.” The Court
in sentence said with dignity, “It is not meet that men should mock
and scoff at persons appointed to any office, yea a necessary office.”
He won one important suit for the town of Gravesend, by which the
right of that town to the entire region of Coney Island was
established; and he received in payment for his legal services
therein, the munificent sum of twenty-four florins (ten dollars) paid in
gray pease. He kept a tavern and was complained of for tapping
after nine o’clock; and he was sued by his landlord for rent; and he
had a yacht, “The Pear Tree,” which ran on trading trips to Albany,
and there were two or three lawsuits in regard to that. He was also a
farmer of the excise on slaughtered cattle; but, in spite of all his
energy and variety of employment, he died insolvent in 1664. The
last lawsuit in which Lawyer Solomon had any share was through a
posthumous connection,—the burgher who furnished an anker of
French wine for the notary’s funeral claimed a position as preferred
creditor to the estate.
A very aggravated case of scorn and resistance of authority was
that of Abel Hardenbrock against the schout de Mill. And this case
shows equally the popular horror of violations of the law and the
confiding trust of the justices that the word of the law was enough
without any visible restraining force. Hardenbrock, who was a
troublesome fellow, had behaved most vilely, shoving the schout on
the breast, and wickedly “wishing the devil might break his neck,”
simply because the schout went to Hardenbrock’s house to warn his
wife not to annoy further Burgomaster De Peyster by unwelcome
visits. Hardenbrock was accordingly seized and made a prisoner at
the Stadt Huys “in the chamber of Pieter Schaefbanck, where he
carried on and made a racket like one possessed and mad,
notwithstanding the efforts of Heer Burgomaster Van Brught, running
up to the Court room and going away next morning as if he had not
been imprisoned.” It was said with amusing simplicity that this cool
walking out of prison was “contrary to the customs of the law,” and a
fine of twenty-five florins was imposed.
For serious words against the government, which could be
regarded as treasonable, the decreed punishment was death. One
Claerbout van ter Goes used such words (unfortunately they are not
given in the indictment), and a judgment was recorded from each
burgomaster and schepen as to what punishment would be proper.
He was branded, whipped on a half-gallows, and banished, and
escaped hanging only by one vote.
All classes in the community were parties in these petty slander
suits; schoolmasters and parsons appear to have been specially
active. Domine Bogardus and Domine Schaets had many a slander
suit. The most famous and amusing of all these clerical suits is the
one brought by Domine Bogardus and his wife, the posthumously
famous Anneke Jans, against Grietje von Salee, a woman of very
dingy reputation, who told in New Amsterdam that the domine’s wife,
Mistress Anneke, had lifted her petticoats in unseemly and extreme
fashion when crossing a muddy street. This was proved to be false,
and the evidence adduced was so destructive of Grietje’s character
that she stands disgraced forever in history as the worst woman in
New Netherland.
Not only were slanderers punished, but they were disgraced with
terrible names. William Bakker was called “a blasphemer, a street
schold, a murderer as far as his intentions are concerned, a defamer,
a disturber of public peace,”—the concentration of which must have
made William Bakker hang his head in the place of his banishment.
They were also rebuked from the pulpit, and admonished in private.
Perhaps the best rebuke given, as well as a unique one, was the
one adopted by Domine Frelinghuysen, who had suffered somewhat
from slander himself. He had this rhyme painted in large letters on
the back of his sleigh, that he who followed might read:—

“Niemands tong; nog neimands pen,

Maakt my anders dan ik ben.
Spreek quaad-spreekers: spreek vonder end,
Niemand en word van u geschend.”

Which, translated into English, reads:—

 “No one’s tongue, and no one’s pen
Makes me other than I am.
Speak, evil-speakers, speak without end,
No one heeds a word you say.”

The original Court of the colony was composed of a Director and

his Council. In 1656, in answer to complaints from the colonists, the
States-General ordered the election of a board of magistrates, in
name and function like those of the Fatherland; namely, a schout,
two burgomasters, and five schepens. The duties of the
burgomasters and schepens were twofold: they regulated municipal
affairs like a board of aldermen, and they sat as a court of justice
both in civil and criminal cases. The annual salary of a burgomaster
was fixed at one hundred and forty dollars, and of a schepen at one
hundred dollars; but as these salaries were to come out of the
municipal chest, which was chronically empty, they never were paid.
When funds did come in from the excise on taverns, on slaughtered
cattle, the tax on land, the fees on transfers, etc., it always had to be
paid out in other ways,—for repairs for the school-room, the Graft,
the watch-room, the Stadt Huys. It never entered the minds of those
guileless civic rulers, two centuries ago, to pay themselves and let
the other creditors go without. The early city schout was also schout-
fiskaal till 1660; but the proper duties of this functionary were really a
combination of those pertaining now to the mayor, sheriff, and
district-attorney. In the little town one man could readily perform all
these duties. He also presided in Court. An offender could thus be
arrested, prosecuted, and judged, by one and the same person,
which seems to us scarcely judicious; but the bench of magistrates
had one useful power, that of mitigating and altering the sentence
demanded by the schout. Often a fine of one hundred guilders would
be reduced to twenty-five; often the order for whipping would be set
aside, and the command of branding as well.
Sometimes justice in New York was tempered with mercy, and
sorely it needed it when fierce English rule and law came in force.
Felons were few, but these few were severely punished. A record of
a trial in 1676 reveals a curious scene in Court, as well as an
astonishing celerity in the execution of the law under English rule
and in the English army. Three soldiers stole a couple of iron pots,
two hoes, a pair of shears, and half a firkin of soap. They were tried
in the morning, confessed, cast into “the Hole” in the afternoon, and
in the evening “the Governor ordered some persons to go to the
prisoners and advise them to prepare for another world, for that one
of them should dye the next day.” On the gloomy morrow, on
Saturday, the three terror-stricken souls drew lots, and the fatal lot
fell to one Thomas Weale. The court of aldermen interceded for him
and finally secured his reprieve till Monday. The peaceful Dutch
Sunday, darkened and shocked by this impending death, saw a
strange and touching sight.
“In the evening a company of the chiefe women of the City,
both English and Dutch, made earnest suite to the Governor
for the Condemned man’s life. Monday in the morning the
same women who came the last night with many others of the
better sort, and a greater number of the ordinary Dutch
women, did again very much importune the Governor to spare
him.”
These tender-hearted colonists were indorsed and supplemented
by the petition of Weale’s fellow-soldiers in the garrison, who
pleaded the prisoner’s youth and his past usefulness, and who
promised if he were pardoned never to steal nor to conceal theft. As
a result of all this intercession, the Governor “graciously” granted
pardon.
This promise and pardon seem to have accomplished much in
army discipline, for thereafter arrests for crime among the soldiery
were rare. Five years later a soldier was accused of pilfering.
“The Court Marshall doth adjudge that the said Melchoir
Classen shall run the Gantlope once, the length of the fort:
where according to the custome of that punishment, the
souldiers shall have switches delivered to them, with which
they shall strike him as he passes between them stript to the
waist, and at the Fort-gate the Marshall is to receive him, and
there to kick him out of the Garrison as a cashiered person,
 when he is no more to returne, and if any pay is due him it is
to be forfeited.”
And that was the end of Melchoir Classen.
Gantlope was the earlier and more correct form of the word now
commonly called gantlet. Running the gantlope was a military
punishment in universal use.
Another common punishment for soldiers (usually for rioting or
drinking) was riding the wooden horse. In New Amsterdam the
wooden horse stood between Paerel Street and the Fort, and was
twelve feet high. Garret Segersen, for stealing chickens, rode the
wooden horse for three days from two o’clock to close of parade with
a fifty-pound weight tied to each foot. At other times a musket was
tied to each foot of the disgraced man. One culprit rode with an
empty scabbard in one hand and a pitcher in the other to show his
inordinate love for John Barleycorn. Jan Alleman, a Dutch officer,
challenged Jan de Fries, who was bedridden; for this cruel and
meaningless insult he too rode the wooden horse. In Revolutionary
days we still find the soldiers of the Continental Army punished by
riding the wooden horse, or, as it was sometimes called, “the timber
mare;” but it was probably a modification of the cruel punishment of
the seventeenth century.
A sailor, for drawing a knife on a companion, was dropped three
times from the yard-arm and received a kick from every sailor on the
ship,—a form of running the gantlope. And we read of a woman who
enlisted as a seaman, and whose sex was detected, being dropped
three times from the yard-arm and tarred and feathered.
These women petitioners for Soldier Weale of whom I have told,
were not the only tender-hearted New Yorkers to petition for “mercy,
that herb of grace, to flower.” During Stuyvesant’s rule his sister,
Madam Bayard, successfully interceded for the release, and thereby
saved the life, of an imprisoned Quaker; and in September, 1713,
two counterfeiters were saved from the death penalty by the
intervention of New York dames. We read, “Most of the gentlewomen
of the city waited on the Governor, and addressed him earnestly with
prayers and tears for the lives of the culprits, who were accordingly
pardoned.” When two sailors rioted through the town demanding
food and drink, and used Carel Van Brugh so roughly that his face
was cut, they were sentenced to be fastened to the whipping-post,
and scourged, and have gashes cut in their faces; the wife of Van
Brugh and her friends petitioned that the sentence should not be
carried out, or at any rate executed within a room. Doubtless other
examples could be found.
The laws of New Netherland were naturally based upon the laws
and customs of the Fatherland, which in turn were formed by the
rules of the College of XIX. from the Imperial Statutes of Charles V.
and the Roman civil law.
The punishments were the ordinary ones of the times, neither
more nor less severe than those of the Fatherland or the other
colonies. In 1691 it was ordered that a ducking-stool be erected in
New York on the wharf in front of the City Hall. The following year an
order was passed that a pillory, cage, and ducking-stool be built.
Though scolds were punished, I have never seen any sentence to
show that this ducking-stool was ever built, or that one was ever
used in New York; while instances of the use of a ducking-stool are
comparatively plentiful in the Southern colonies. The ducking-stool
was an English “engine” of punishment, not a Dutch.
The colonists were astonishingly honest. Thieves were surprisingly
few; they were punished under Dutch rule by scourging with rods,
and usually by banishment,—a very convenient way of shifting
responsibility. Assaults were punished by imprisonment and
subjection to prison fare, consisting only of bread and water or small
beer; and sometimes temporary banishment. There was at first no
prison, so men were often imprisoned in their own houses, which
does not seem very disgraceful. In the case of François de Bruyn,
tried for insulting and striking the court messenger, he was fined two
hundred guilders, and answered that he would rot in prison before he
would pay. He was then ordered to be imprisoned in a respectable
tavern, which sentence seems to have some possibility of mitigating
accompaniments.
 In 1692 it was ordered in Kings County that a good pair of stocks
and a pound be made in every bound within Kings County, and kept
in sufficient repair. In repair and in use were they kept till this century.
Pillories too were employed in punishment till within the memory of
persons now living. The whipping-post was really a public blessing,
—in constant use, and apparently of constant benefit, though the
publicity of its employment seems shocking to us to-day. The public
whipper received a large salary. In 1751, we learn from an
advertisement, it was twenty pounds annually.
Some of the punishments were really almost picturesque in their
ingenious inventions of mortification and degradation. Truly it was a
striking sight when “Jan of Leyden”—a foul-mouthed rogue, a true
blather-schuyten—was fastened to a stake in front of the townhouse,
with a bridle in his mouth and a bundle of rods tied under each arm,
and a placard on his breast bearing the inscription, “Lampoon-riter,
false accuser, defamer of magistrates.” Though he was banished, I
am sure he never was forgotten by the children who saw him
standing thus garnished and branded on that spring day in 1664. In
the same place a thief was punished by being forced to stand all day
under a gallows, a gallows-rope around his neck and empty sword-
scabbard in his hand, a memorable figure.
And could any who saw it ever forget the punishment of Mesaack
Martens, who stole six cabbages from his neighbor, and confessed
and stood for days in the pillory with cabbages on his head, that “the
punishment might fit the crime;” to us also memorable because the
prisoner was bootlessly examined by torture to force confession of
stealing fowls, butter, turkeys, etc.
He was not the only poor creature who suffered torture in New
Amsterdam. It was frequently threatened and several times
executed. The mate of a ship was accused of assaulting a sheriff’s
officer, who could not identify positively his assailant. The poor mate
was put to torture, and he was innocent of the offence. The assailant
was proved to be another man from whom the officer had seized a
keg of brandy. Still none in New Amsterdam were tortured or pressed
to death. The blood of no Giles Corey stains the honor of New
Netherland.
 Sometimes the execution of justice seemed to “set a thief to catch
a thief.” A letter written by an English officer from Fort James on
Manhattan Island to Captain Silvester Salisbury in Fort Albany in
1672 contains this sentence:—
“We had like to have lost our Hang-man Ben Johnson, for
he being taken in Divers Thefts and Robbings convicted and
found guilty, escaped his neck through want of another
Hangman to truss him up, soe that all the punishment that he
received for his Three Years’ Roguery in thieving and stealing
(which was never found out till now) was only 39 stripes at the
Whipping Post, loss of an Ear and Banishment.”
We have the records of an attempt at capital punishment in 1641;
and Mr. Gerard’s account of it in his paper “The Old Stadt-Huys” is
so graphic, I wish to give it in full:—
“The court proceedings before the Council, urged by the
Fiscal, were against Jan of Fort Orange, Manuel Gerrit the
Giant, Anthony Portugese, Simon Congo, and five others, all
negroes belonging to the Company, for killing Jan Premero,
another negro. The prisoners having pleaded guilty, and it
being rather a costly operation to hang nine able-bodied
negroes belonging to the Company, the sentence was that
they were to draw lots to determine ‘who should be punished
with the cord until death, praying the Almighty God, the
Creator of Heaven and Earth, to direct that the lot may fall on
the guiltiest, whereupon’ the record reads, ‘the lot fell by
God’s Providence on Manuel Gerrit, the Giant, who was
accordingly sentenced to be hanged by the neck until dead as
an example to all such malefactors.’ Four days after the trial,
and on the day of the sentence, all Nieuw Amsterdam left its
accustomed work to gaze on the unwonted spectacle. Various
Indians also gathered, wondering, to the scene. The giant
negro is brought out by the black hangman, and placed on the
ladder against the fort with two strong halters around his
neck. After an exhortation from Domine Bogardus during
which the negro chaunts barbaric invocations to his favorite
Fetich, he is duly turned off the ladder into the air. Under the
 violent struggles and weight of the giant, however, both
halters break. He falls to the ground. He utters piteous cries.
Now on his knees, now twisting and groveling on the earth.
The women shriek. The men join in his prayers for mercy to
the stern Director. He is no trifler and the law must have its
course. The hangman prepares a stronger rope. Finally the
cry for mercy is so general that the Director relents, and the
fortunate giant is led off the ground by his swarthy friends,
somewhat disturbed in his intellect by his near view of the
grim King of Terrors.”
Up to February 21, 1788, benefit of clergy existed; that is, the plea
in capital felonies of being able to read. This was a monkish privilege
first extended only to priestly persons. In England it was not
abolished till 1827. The minutes of the Court of General Quarter
Sessions in New York bear records of criminals who pleaded “the
benefit” and were branded on the brawn of the left thumb with “T” in
open court and then discharged.
As the punishments accorded for crimes were not severe for the
notions of the times, it is almost amusing to read some fierce
ordinances,—though there is no record of any executions in
accordance with them. For instance, in January, 1659, by the
Director-General and Council with the advice of the burgomasters
and schepens it was enacted that “No person shall strip the fences
of posts or rails under penalty for the first offence of being whipped
and branded, and for the second, of punishment with the cord until
death ensues.” It is really astonishing to think of these kindly Dutch
gentlemen calmly ordering hanging for stealing fence-rails, though of
course the matter reached further than at first appeared: there was
danger of a scarcity of grain; and if the fences were stolen, the cattle
would trample down and destroy the grain. Later orders as to fences
were given which appear eminently calculated to be mischief-
making. “Persons thinking their neighbors’ fences not good, first to
request them to repair; failing which to report to the overseers.” In
1674 all persons were forbidden to leave the city except by city-gate,
under penalty of death; this was of course when war threatened.
 The crime of suicide was not without punishment. Suicides were
denied ordinary burial rites. In Dutch days when one Smitt of New
York committed suicide, the schout asked that his body be drawn on
a hurdle and buried with a stake in his heart. This order was not
executed; he was buried at night and his estates confiscated. When
Sir Danvers Osborne—the Governor for a day—was found dead by
his own act, he was “decently interred in Trinity churchyard.”
Women in New York sometimes made their appearance in New
York courts, as in those of other colonies, in another rôle than that of
witness or criminal; they sometimes sat on juries. In the year 1701,
six good Albany wives served on a jury: Tryntje Roseboom, Catheren
Gysbertse, Angeneutt Jacobse, Marritje Dirkse, Elsje Lansing, and
Susanna Bratt. They were, of course, empanelled for a special duty,
not to serve on the entire evidence of the case for which they were
engaged.
Many old records are found which employ quaint metaphors or
legal expressions; I give one which refers to a custom which seems
at one time to have been literally performed. It occurs in a
commission granted to the trustees of an estate of which the debts
exceeded the assets. Any widow in Holland or New Netherland could
be relieved of all demands or claims of her husband’s creditors by
relinquishing all right of inheritance. This widow took this privilege; it
is recorded thus:—
“Whereas, Harman Jacobsen Bamboes has been lately
shot dead, murdered by the Indians, and whereas the estate
left by him has been kicked away with the foot by his wife who
has laid the key on the coffin, it is therefore necessary to
authorize and qualify some persons to regulate the same.”
There was a well-known Dutch saying which referred to this
privilege, Den Sleutel op het graf leggen, and simply meant not to
pay the debts of the deceased.
This legal term and custom is of ancient origin. In Davies’ “History
of Holland” we read of a similar form being gone through with in
Holland in 1404, according to the law of Rhynland. The widow of a
great nobleman immediately after his death desired to renounce all
claim on his estate and responsibility for his debts. She chose a
guardian, and, advancing with him to the door of the Court (where
the body of the dead Count had been placed on a bier), announced
that she was dressed wholly in borrowed clothing; she then formally
gave a straw to her guardian, who threw it on the dead body, saying
he renounced for her all right of dower, and abjured all debts. This
was derived from a still more ancient custom of the Franks, who
renounced all alliances by the symbolic breaking and throwing away
a straw.
In other states of the Netherlands the widow gave up dower and
debts by laying a key and purse on the coffin. This immunity was
claimed by persons in high rank, one being the widow of the Count
of Flanders.
In New England (as I have told at length in my book, “Customs
and Fashions in Old New England,”) the widow who wished to
renounce her husband’s debts was married in her shift, often at the
cross-roads, at midnight. These shift-marriages took place in
Massachusetts as late as 1836; I have a copy of a court record of
that date.
I know of but one instance of the odious and degrading English
custom of wife-trading taking place in New York. Laurens Duyts, an
agent for Anneke Jans in some of her business transactions, was in
the year 1663 sentenced to be flogged and have his right ear cut off
for selling his wife, Mistress Duyts, to one Jansen. Possibly the
severity of the punishment may have prevented the recurrence of the
crime.
After a somewhat extended study and comparison of the early
court and church records of New England with those of New York, I
cannot fail to draw the conclusion—if it is just to judge from such
comparisons—that the state of social morals was higher in the Dutch
colonies than in the English. Perhaps the settlers of Boston and
Plymouth were more severe towards suspicion of immorality, as they
were infinitely more severe towards suspicion of irreligion, than were
their Dutch neighbors. And they may have given more publicity and
punishment to deviations from the path of rectitude and uprightness;
but certainly from their own records no fair-minded person can fail to
deem them more frail, more erring, more wicked, than the Dutch.
The circumstances of immigration and the tendencies of
temperament were diverse, and perhaps it was natural that a
reaction tending to sin and vice should come to the intense and
overwrought religionist rather than to the phlegmatic and prosperous
trader. In Virginia and Maryland the presence of many convict-
emigrants would form a reasonable basis for the existence of the
crime and law-breaking which certainly was in those colonies far in
excess of the crime in New Netherland and New York.
I know that Rev. Mr. Miller, the English clergyman, did not give the
settlement a very good name at the last of the seventeenth century;
but even his strictures cannot force me to believe the colonists so
unbearably wicked.
It should also be emphasized that New Netherland was far more
tolerant, more generous than New England to all of differing religious
faiths. Under Stuyvesant, however, Quakers were interdicted from
preaching, were banished, and one Friend was treated with great
cruelty. The Dutch clergymen opposed the establishment of a
Lutheran church, and were rebuked by the Directors in Holland, who
said that in the future they would send out clergymen “not tainted
with any needless preciseness;” and Stuyvesant was also rebuked
for issuing an ordinance imposing a penalty for holding conventicles
not in accordance with the Synod of Dort. Many Christians not in
accordance in belief with that synod settled in New Netherland.
Quakers, Lutherans, Church of England folk, Anabaptists,
Huguenots, Waldenses, Walloons. The Jews were protected and
admitted to the rights of citizenship. Director Kieft, with heavy
ransoms, rescued the captive Jesuits, Father Jogues and Father
Bressani, from the Indians and tenderly cared for them. No witches
suffered death in New York, and no statute law existed against
witchcraft. There is record of but one witchcraft trial under the
English governor, Nicholls, who speedily joined with the Dutch in
setting aside all that nonsense.
 CHAPTER XIII
CHURCH AND SUNDAY IN OLD NEW YORK

Sunday was not observed in New Netherland with any such rigidity
as in New England. The followers of Cocceius would not willingly
include Saturday night, and not even all of the Sabbath day, in their
holy time. Madam Knight, writing in 1704 of a visit to New York,
noted: “The Dutch aren’t strict in keeping the Sabbath as in Boston
and other places where I have been.” This was, of course, in times of
English rule in New York. Still, much respect to the day was required,
especially under the governing hand of the rigid Calvinist
Stuyvesant. He specially enjoined and enforced strict regard for
seemly quiet during service time. The records of Stuyvesant’s
government are full of injunctions and laws prohibiting “tavern-
tapping” during the hours of church service. He would not tolerate
fishing, gathering of berries or nuts, playing in the street, nor gaming
at ball or bowls during church time. At a little later date the time of
prohibition of noise and tapping and gaming was extended to include
the entire Sabbath day, and the schout was ordered to be active in
searching out and punishing such offenders.
Occasionally his vigilance did discover some Sabbath disorders.
He found the first Jew trader who came to the island of Manhattan
serenely keeping open shop on Sunday, and selling during sermon
time, knowing naught of any Sunday laws of New Amsterdam.
And Albert the Trumpeter was seen on the Sabbath in suspicious
guise, with an axe on his shoulder,—but he was only going to cut a
bat for his little son; and as for his neighbor who did cut wood, it was
only kindling, since his children were cold.
And one Sunday evening in 1660 the schout triumphantly found
three sailors round a tap-house table with a lighted candle and a
backgammon-board thereon; and he surely had a right to draw an
inference of gaming therefrom.
And in another public-house ninepins were visible, and a can and
glass, during preaching-time. The landlady had her excuse,—some
came to her house and said church was out, and one chanced to
have a bowl in his hand and another a pin, but there was no playing
at bowls.
Still, though he snooped and fined, in 1656 the burgomasters
learned “by daily and painful experience” that the profanation of “the
Lord’s day of Rest by the dangerous, Yes, damnable Sale or Dealing
out of Wines Beers and Brandy-Waters” still went on; and fresh
Sunday Laws were issued forbidding “the ordinary and customary
Labors of callings, such as Sowing, Mowing, Building, Sawing wood,
Smithing, Bleeching, Hunting, Fishing.” All idle sports were banned
and named: “Dancing, Card-playing, Tick-tacking, Playing at ball, at
bowls, at ninepins; taking Jaunts in Boats, Wagons, or Carriages.”
In 1673, again, the magistrates “experienced to our great grief”
that rolling ninepins was more in vogue on Sunday than on any other
day. And we learn that there were social clubs that “Set on the
Sabbath,” which must speedily be put an end to. Thirty men were
found by the schout in one tap-huys; but as they were playing
ninepins and backgammon two hours after the church-doors had
closed, prosecution was most reluctantly abandoned.
Of course scores of “tappers” were prosecuted, both in taverns
and private houses. Piety and regard for an orderly Sabbath were
not the only guiding thoughts in the burgomasters’ minds in framing
these Sunday liquor laws and enforcing them; for some tapsters had
“tapped beer during divine service and used a small kind of measure
which is in contempt of our religion and must ruin our state,”—and
the state was sacred. In the country, as for instance on Long Island,
the carting of grain, travelling for pleasure, and shooting of wild-fowl
on Sunday were duly punished in the local courts.
I do not think that children were as rigid church attendants in New
York as in New England. In 1696, in Albany, we find this injunction:
“ye Constables in eache warde to take thought in attending at ye
church to hender such children as Profane ye Sabbath;” and we
know that Albany boys and girls were complained of for coasting
down hill on Sunday,—which enormity would have been simply
impossible in New England, except in an isolated outburst of Adamic
depravity. In another New York town the “Athoatys” complained of
the violation of the Sabbath by “the Younger Sort of people in
Discourssing of Vane things and Running of Raesses.” As for the city
of New York, even at Revolutionary times a cage was set up on City
Hall Park in which to confine wicked New York boys who profaned
the Sabbath. I do not find so full provisions made for seating children
in Dutch Reformed churches as in Puritan meeting-houses. A wise
saying of Martin Luther’s was “Public sermons do very little edify
children”—perhaps the Dutch agreed with him. As the children were
taught the Bible and the catechism every day in the week, their
spiritual and religious schooling was sufficient without the Sunday
sermon,—but, of course, if they were not in the church during
services, they would “talk of vane Things and run Raesses.”
Before the arrival of any Dutch preacher in the new settlement in
the new world, the spiritual care of the little company was provided
for by men appointed to a benign and beautiful old Dutch office, and
called krankebesoeckers or zeikentroosters,—“comforters of the
Sick,”—who not only tenderly comforted the sick and weary of heart,
but “read to the Commonalty on Sundays from texts of Scripture with
the Comments.” These pious men were assigned to this godly work
in Fort Orange and in New Amsterdam and Breuckelen. In Esopus
they had meetings every Sunday, “and one among us read
something for a postille.” Often special books of sermons were read
to the congregations.
In Fort Orange they had a domine before they had a church. The
patroon instructed Van Curler to build a church in 1642; but it was
not until 1646 that the little wooden edifice was really put up. It was
furnished at a cost of about thirty-two dollars by carpenter
Fredricksen, with a predickstoel, or pulpit, a seat for the magistrates,
—de Heerebanke,—one for the deacons, nine benches and several
corner-seats.