Knowledge and Systems Sciences 19th

International Symposium KSS 2018

Tokyo Japan November 25 27 2018
Proceedings Jian Chen
Visit to download the full and correct content document:
https://textbookfull.com/product/knowledge-and-systems-sciences-19th-international-
symposium-kss-2018-tokyo-japan-november-25-27-2018-proceedings-jian-chen/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Knowledge and Systems Sciences 17th International

Symposium KSS 2016 Kobe Japan November 4 6 2016
Proceedings 1st Edition Jian Chen

https://textbookfull.com/product/knowledge-and-systems-
sciences-17th-international-symposium-kss-2016-kobe-japan-
november-4-6-2016-proceedings-1st-edition-jian-chen/

PRIMA 2018 Principles and Practice of Multi Agent

Systems 21st International Conference Tokyo Japan
October 29 November 2 2018 Proceedings Tim Miller

https://textbookfull.com/product/prima-2018-principles-and-
practice-of-multi-agent-systems-21st-international-conference-
tokyo-japan-october-29-november-2-2018-proceedings-tim-miller/

Smart Blockchain First International Conference

SmartBlock 2018 Tokyo Japan December 10 12 2018
Proceedings Meikang Qiu

https://textbookfull.com/product/smart-blockchain-first-
international-conference-smartblock-2018-tokyo-japan-
december-10-12-2018-proceedings-meikang-qiu/

Smart Computing and Communication Third International

Conference SmartCom 2018 Tokyo Japan December 10 12
2018 Proceedings Meikang Qiu

https://textbookfull.com/product/smart-computing-and-
communication-third-international-conference-smartcom-2018-tokyo-
japan-december-10-12-2018-proceedings-meikang-qiu/
Advanced Concepts for Intelligent Vision Systems 19th
International Conference ACIVS 2018 Poitiers France
September 24 27 2018 Proceedings Jacques Blanc-Talon

https://textbookfull.com/product/advanced-concepts-for-
intelligent-vision-systems-19th-international-conference-
acivs-2018-poitiers-france-september-24-27-2018-proceedings-
jacques-blanc-talon/

Web Information Systems Engineering – WISE 2018: 19th

International Conference, Dubai, United Arab Emirates,
November 12-15, 2018, Proceedings, Part II Hakim Hacid

https://textbookfull.com/product/web-information-systems-
engineering-wise-2018-19th-international-conference-dubai-united-
arab-emirates-november-12-15-2018-proceedings-part-ii-hakim-
hacid/

Knowledge Engineering and Knowledge Management 21st

International Conference EKAW 2018 Nancy France
November 12 16 2018 Proceedings Catherine Faron Zucker

https://textbookfull.com/product/knowledge-engineering-and-
knowledge-management-21st-international-conference-
ekaw-2018-nancy-france-november-12-16-2018-proceedings-catherine-
faron-zucker/

Security Standardisation Research 4th International

Conference SSR 2018 Darmstadt Germany November 26 27
2018 Proceedings Cas Cremers

https://textbookfull.com/product/security-standardisation-
research-4th-international-conference-ssr-2018-darmstadt-germany-
november-26-27-2018-proceedings-cas-cremers/

Semantic Technology 8th Joint International Conference

JIST 2018 Awaji Japan November 26 28 2018 Proceedings
Ryutaro Ichise

https://textbookfull.com/product/semantic-technology-8th-joint-
international-conference-jist-2018-awaji-japan-
november-26-28-2018-proceedings-ryutaro-ichise/
Jian Chen · Yuji Yamada
Mina Ryoke · Xijin Tang (Eds.)

Communications in Computer and Information Science 949

Knowledge and
Systems Sciences
19th International Symposium, KSS 2018
Tokyo, Japan, November 25–27, 2018
Proceedings

123
Communications
in Computer and Information Science 949
Commenced Publication in 2007
Founding and Former Series Editors:
Phoebe Chen, Alfredo Cuzzocrea, Xiaoyong Du, Orhun Kara, Ting Liu,
Dominik Ślęzak, and Xiaokang Yang

Editorial Board
Simone Diniz Junqueira Barbosa
Pontifical Catholic University of Rio de Janeiro (PUC-Rio),
Rio de Janeiro, Brazil
Joaquim Filipe
Polytechnic Institute of Setúbal, Setúbal, Portugal
Ashish Ghosh
Indian Statistical Institute, Kolkata, India
Igor Kotenko
St. Petersburg Institute for Informatics and Automation of the Russian
Academy of Sciences, St. Petersburg, Russia
Krishna M. Sivalingam
Indian Institute of Technology Madras, Chennai, India
Takashi Washio
Osaka University, Osaka, Japan
Junsong Yuan
University at Buffalo, The State University of New York, Buffalo, USA
Lizhu Zhou
Tsinghua University, Beijing, China
More information about this series at http://www.springer.com/series/7899
Jian Chen Yuji Yamada
•

Mina Ryoke Xijin Tang (Eds.)

•

Knowledge and
Systems Sciences
19th International Symposium, KSS 2018
Tokyo, Japan, November 25–27, 2018
Proceedings

123
Editors
Jian Chen Mina Ryoke
School of Economics and Management Faculty of Business Sciences
Tsinghua University University of Tsukuba
Beijing, China Tokyo, Japan
Yuji Yamada Xijin Tang
Graduate School of Business Sciences Institute of Systems Science
University of Tsukuba CAS Academy of Mathematics and Systems
Bunkyō, Tokyo, Tokyo, Japan Sciences
Beijing, China

ISSN 1865-0929 ISSN 1865-0937 (electronic)

Communications in Computer and Information Science
ISBN 978-981-13-3148-0 ISBN 978-981-13-3149-7 (eBook)
https://doi.org/10.1007/978-981-13-3149-7

Library of Congress Control Number: 2018960676

© Springer Nature Singapore Pte Ltd. 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 Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
 Preface

The annual International Symposium on Knowledge and Systems Sciences aims to

promote the exchange and interaction of knowledge across disciplines and borders to
explore the new territories and new frontiers. With over 18-year continuous endeavors,
attempts to strictly define knowledge science may be still ambitious, but a very tolerant,
broad-based, and open-minded approach to the discipline can be taken. Knowledge
science and systems science can complement and benefit each other methodologically.
The First International Symposium on Knowledge and Systems Sciences
(KSS 2000) was initiated and organized by the Japan Advanced Institute of Science and
Technology (JAIST) in September of 2000. Since then KSS 2001 (Dalian),
KSS 2002 (Shanghai), KSS 2003 (Guangzhou), KSS 2004 (JAIST), KSS 2005
(Vienna), KSS 2006 (Beijing), KSS 2007 (JAIST), KSS 2008 (Guangzhou), KSS 2009
(Hong Kong), KSS 2010 (Xi’an), KSS 2011 (Hull), KSS 2012 (JAIST), KSS 2013
(Ningbo), KSS 2014 (Sapporo), KSS 2015 (Xi’an), KSS 2016 (Kobe), and KSS 2017
(Bangkok) have been a successful platform for many scientists and researchers from
different countries. During the past 18 years, people interested in knowledge and sys-
tems sciences have become a community, and an international academic society has
existed for 15 years.
This year KSS was held in Tokyo, Japan, during November 25–27, 2018. The
conference provided opportunities for presenting interesting new research results and
facilitating interdisciplinary discussions, leading to knowledge transfer under the theme
of “Knowledge Acquisition from Structured and Unstructured Data for Effective Social
Implementation.” To fit that theme, four distinguished scholars delivered the keynote
speeches.
• Chonghui Guo (Dalian University of Technology, China), “Big Data Analytics in
Health Care: Data driven Methods for typical Diagnosis and Treatment Pattern
Mining”
• Setsuya Kurahashi (University of Tsukuba, Japan), “Model-Based Policy Making:
Urban Dynamics, Collaborative Learning, and Family Strategy”
• Yoichi Motomura (Artificial Intelligence Research Center, National Institute of
Advanced Industrial Science and Technology, Japan), “Toward Cyber Physical
Innovation: Probabilistic Modeling for Real-Field AI Applications”
• Thanaruk Theeramunkong (Sirindhorn International Institute of Technology,
Thammasat University, Thailand), “Text Mining from Public Hearing Databases
and Automatic Profile Generation from Online Resources”
KSS 2018 received 54 submissions from authors studying and working in Belgium,
China, India, Indonesia, Japan, Sri Lanka, Thailand, and Russia, and finally 20 sub-
missions were selected for publication in the proceedings after a double-blind review
process. The co-chairs of the international Program Committee made the final decision
VI Preface

for each submission based on the review reports from the referees, who came from
China, Japan, New Zealand, Thailand, and the USA.
We received a lot of support and help from many people and organizations. We
would like to express our sincere thanks to the authors for their remarkable contri-
butions, all the Technical Program Committee members for their time and expertise in
reviewing the papers within a very tight schedule, and the proceedings publisher
Springer for their professional help. It is the third time that the KSS proceedings are
published as a CCIS volume after successful collaboration with Springer in 2016 and
2017. We greatly appreciate our four distinguished scholars for accepting our invitation
to deliver keynote speeches at the symposium. Last but not least, we are very indebted
to the local organizers for their hard work.
We were happy to witness the thought-provoking and lively scientific exchanges in
the fields of knowledge and systems sciences during the symposium.

November 2018 Jian Chen

Yuji Yamada
Mina Ryoke
Xijin Tang
 Organization

Organizer

International Society for Knowledge and Systems Sciences

Host

University of Tsukuba, Tokyo Campus, Japan

General Chairs
Jian Chen Tsinghua University, China
Yuji Yamada University of Tsukuba, Japan

Program Committee Chairs

Mina Ryoke University of Tsukuba, Japan
Xijin Tang CAS Academy of Mathematics and Systems Science,
China
Jiangning Wu Dalian University of Technology, China

Technical Program Committee

Quan Bai Auckland University of Technology, New Zealand
Masataka Ban University of Tsukuba, Japan
Meng Cai Xidian University, China
Zhigang Cao Beijing Jiaotong University, China
Hao Chen Nankai Univertsity, China
Jindong Chen Beijing Information Science and Technology University,
China
Zengru Di Beijing Normal University, China
Yong Fang CAS Academy of Mathematics and Systems Science,
China
Chonghui Guo Dalian University of Technology, China
Jun Huang Angelo State University, USA
Van-Nam Huynh Japan Advanced Institute of Science and Technology,
Japan
Setsuya Kurahashi University of Tsukuba, Japan
Bin Jia Beijing Jiaotong University, China
Jiradett Kerdsri Defense Technology Institute, Thailand
Weidong Li Shaanxi Normal University, China
VIII Organization

Yi Li Xidian University, China

Zhenpeng Li Dali University, China
Jiamou Liu University of Auckland, New Zealand
Yijun Liu CAS Institute of Science and Development, China
Patiyuth Pramkeaw King Mongkut’s University of Technology Thonburi,
Thailand
Mina Ryoke University of Tsukuba, Japan
Hidenori Sato University of Tsukuba, Japan
Tadahiko Sato University of Tsukuba, Japan
Bingzhen Sun Xidian University, China
Xijin Tang CAS Academy of Mathematics and Systems Science,
China
Jing Tian Wuhan University of Technology, China
Haibo Wang Texas A&M International University, USA
Cuiping Wei Yangzhou University, China
Jiang Wu Wuhan University, China
Jiangning Wu Dalian University of Technology, China
Jinpeng Xu Xidian University, China
Thaweesak King Mongkut’s University of Technology Thonburi,
Yingthawornsuk Thailand
Kenichi Yoshida University of Tsukuba, Japan
Wen Zhang Beijing University of Technology, China
Zhen Zhang Dalian University of Technology, China
Abstracts of Keynotes
Big Data Analytics in Healthcare: Data-Driven
Methods for Typical Diagnosis and Treatment
Pattern Mining

Chonghui Guo

Institute of Systems Engineering, Faculty of Management and Economics,

Dalian University of Technology, China
dlutguo@dlut.edu.cn

Abstract. A huge volume of digitized clinical data is generated and accumulated

rapidly since the widespread adoption of Electronic Medical Records (EMRs).
These massive quantities of data hold the promise of propelling healthcare
evolving from a proficiency-based art to a data-driven science, from a reactive
mode to a proactive mode, from one-size-fits-all medicine to personalized
medicine. Personalized medicine refers to tailoring medical diagnosis and
treatment to the individual characteristics of each patient, which literally means
the ability to classify individuals into subpopulations that differ in their sus-
ceptibility to a disease or their response to a specific treatment. While EMRs
contain rich temporal and heterogeneous medical information that can be used
for typical diagnosis and treatment pattern mining by big data analytics. Hence,
this study will analyze different data types of EMRs in depth and design
data-driven EMRs mining method, including data-driven typical diagnosis
pattern extraction from multi-type data of EMRs, and data-driven typical
treatment pattern extraction from multi-view of doctor orders. Specifically, for
typical diagnosis pattern extraction, we first design three similarity measure
methods for patient demographic, symptom, and laboratory examination infor-
mation, then adopt similarity fusion method to generate a unified similarity and
construct similarity network of patient hospital admission, next propose a patient
diagnostic information similarity method by integrating patient hospital
admission information, and finally perform clustering algorithm to extract typ-
ical diagnosis patterns. For typical treatment pattern extraction, we first study
automatic treatment regimen development and recommendation from the con-
tent view of doctor orders, then study typical treatment process extraction and
evaluation from the sequence view of doctor orders, next study typical drug use
pattern extraction and evaluation from the duration view of doctor orders, and
finally propose a fusion framework for typical treatment pattern extraction from
multi-view of doctor orders. Furthermore, all proposed methods have been
validated on real-world EMRs of cerebral infarction dataset and MIMIC-III
dataset.
Model-Based Policy Making: Urban Dynamics,
Collaborative Learning and Family Strategy

Setsuya Kurahashi

Faculty of Business Sciences, University of Tsukuba, Japan

kurahashi.setsuya.gf@u.tsukuba.ac.jp

Abstract. Many significant policies of our society and economy are determined
by someone day after day. However, most of the plans have been discussed and
decided based on past experiences and data. Many of them estimate policy
effects by analyzing actual phenomena and data using statistical methods. For
this method called evidence-based policymaking (EBP), this lecture proposes
model-based policymaking (MBP). The MBP is designed with an agent-based
model and data science techniques, and it also called as social simulation. The
model-based approach enables to design realistic phenomena as a model and
predict the effect on unfolding future events due to hypotheses or activities that
are difficult to experiment using computer experiments. In the field of business
and sociology, data analysis as an induction method and strategy planning as a
deductive method are connected. In the lecture, I will introduce urban dynamics
model, teaching model at school, analysis of education in a family using a
genealogy in China during 500 years.
 Toward Cyber Physical Innovation:
Probabilistic Modeling
for Real Field AI Applications

Yoichi Motomura

Artificial Intelligence Research Center,

National Institute of Advanced Industrial Science and Technology, Japan
y.motomura@aist.go.jp

Abstract. Currently, the practical application of artificial intelligence is dra-

matically advanced by machine learning using real world big data. Industrial
structure reform and the smart society called Society 5.0 are also expected to be
realized. In this talk, real world application and the research projects on AI are
introduced. Our social system is changing into a Cyber Physical System by AI
and real world big data. For example, point of sales data (POS-data) is linked to
the customer ID by the common point card system. Moreover, sensors in a store
can capture a customer’s behavior during shopping process. From phenomena of
real world big data with high temporal and spatial resolution, phenomena can be
represented by a probabilistic model that can be calculated, predicting risk, cost,
benefit and making it possible to simulate. In our system, this computational
process is realized by PLSA (Probabilistic Latent Semantic Analysis) and
Bayesian networks.
We developed interactive digital signage systems and interactive vending
machines driven by AI. These systems are being investigated in the use cases
such as improvement of productivity of services such as management support
and logistics optimization. The same framework can be utilized also child care
and health promotion activities for local community support.
 Text Mining from Public Hearing Databases
and Automatic Profile Generation
from Online Resources

Thanaruk Theeramunkong

Sirindhorn International Institute of Technology,

Thammasat University, Thailand
thanaruk@siit.tu.ac.th

Abstract. Analyzing natural language texts helps us to obtain information or

knowledge for various purposes. This talk firstly provides a short summary
of the state of the art on research and development in natural language pro-
cessing, including language characteristics, rule-based and statistical method-
ologies, as well as difficulties and challenges. Secondly, with the growing
availability of fact-oriented and/or opinion-rich online textual contents, I present
new opportunities and challenges of using text mining techniques to seek out
and understand the facts and the opinions in our society. Thirdly, along with this
trend, electronic public hearing information, incident information and personal
activity logs are described and the potentials towards knowledge discovery are
enumerated and two applications of social monitoring and personal profile
generation are discussed. For the first task, from 2014 to 2015, during the
National Reform Council, there have been an activity of reform-related public
hearing in Thailand. The information was applied for analyzing Thai opinion on
the country’s reform process and topics for reforming. The second task is to
gather online information related to individual activities/events for generating
personal profiles. Such information can be used for characterizing individuals
for expert recruiting or seeking. In the talk, I report the progress of our research
works on these two tasks, including their potential use in the future.

Keywords: Representation learning • Deep learning • COSFIRE trainable filters

 Contents

Modeling the Heterogeneous Mental Accounting Impacts

of Inter-shopping Duration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Kazuhiro Miyatsu and Tadahiko Sato

The Characteristics of Service Efficiency and Patient Flow

in Heavy Load Outpatient Service System . . . . . . . . . . . . . . . . . . . . . . . . . 17
Yuan Xu, Xiaopu Shang, Hongmei Zhao, Runtong Zhang, and Jun Wang

How Do You Reduce Waiting Time? Analytical Model Expansion

with Unstructured Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Keiichi Ueda and Setsuya Kurahashi

When We Talk About Medical Service, What Do We Concern?

A Text Analysis of Weibo Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Ke Wang, Chaocheng He, Lin Wang, and Jiang Wu

Modeling Wicked Problems in Healthcare Using Interactive Qualitative

Analysis: The Case of Patients’ Internet Usage . . . . . . . . . . . . . . . . . . . . . . 58
Renuka Devi S. Karthikeyan, Prakash Sai Lokachari, and Nargis Pervin

Mining Typical Drug Use Patterns Based on Patient Similarity

from Electronic Medical Records . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Jingfeng Chen, Chonghui Guo, Leilei Sun, and Menglin Lu

Aiding First Incident Responders Using a Decision Support System

Based on Live Drone Feeds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Jerico Moeyersons, Pieter-Jan Maenhaut, Filip De Turck,
and Bruno Volckaert

The Use of Event-Based Modeling and System-Dynamics Modeling

in Accident and Disaster Investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Xiangting Chen and Xiao Liu

Generating Risk Maps for Evolution Analysis of Societal Risk Events. . . . . . 115
Nuo Xu and Xijin Tang

Research on Forest Fire Processing Scheme Generation Method

Based on Belief Rule-Base . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Yan Xu, Ning Wang, Xuehua Wang, Zijian Ni, Huaiming Li,
and Xuelong Chen

Kansei Knowledge Extraction as Measure of Structural Heterogeneity . . . . . . 142

Mina Ryoke and Tadahiko Sato
XVI Contents

The Impact of Online Reviews on Product Sales: What’s Role

of Supplemental Reviews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Hao Liu, Jiangning Wu, Xian Yang, and Xianneng Li

Social Media and the Diffusion of an Information Technology Product . . . . . 171

Yinxing Li and Nobuhiko Terui

The Effect of Cognitive Trust on Team Performance:

A Deep Computational Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Deqiang Hu, Yanzhong Dang, Xin Yue, and Guangfei Yang

Link Prediction Based on Supernetwork Model and Attention Mechanism . . . 201

Yuxue Chi and Yijun Liu

Dynamics of Deffuant Model in Activity-Driven Online Social Network . . . . 215

Jun Zhang, Haoxiang Xia, and Peng Li

Identifying Factors that Impact on the Learning Process of Sewing Workers

on an Assembly Line. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Thanh Quynh Song Le and Van Nam Huynh

A Study on Constructing KM System for Laboratories Based

on the Three-Stage EDIS Spiral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Bingfei Tian, Jianwen Xiang, Ming Yang, Dongdong Zhao,
and Jing Tian

Some q-Rung Orthopair Fuzzy Dual Maclaurin Symmetric Mean Operators

with Their Application to Multiple Criteria Decision Making . . . . . . . . . . . . 252
Jun Wang, Runtong Zhang, Li Li, Xiaopu Shang, Weizi Li, and Yuan Xu

An Improved Short Pause Based Voice Activity Detection

Using Long Short-Term Memory Recurrent Neural Network . . . . . . . . . . . . 267
Kiettiphong Manovisut, Pokpong Songmuang,
and Nattanun Thatphithakkul

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275

 Modeling the Heterogeneous Mental
Accounting Impacts of Inter-shopping
Duration

Kazuhiro Miyatsu1(B) and Tadahiko Sato2

1
The Nielsen Company, Tokyo, Japan
kazuhiro.miyatsu@nielsen.com
2
Faculty of Business Sciences, University of Tsukuba, Tokyo, Japan
sato@gssm.otsuka.tsukuba.ac.jp

Abstract. Unlike the principles of traditional economics, namely that

goods with monetary equivalency can be substituted, mental account-
ing states that these goods have different criteria values to consumers
depending on the purposes of their use and circumstances at purchase.
By modeling an inter-shopping duration that accommodates the mental
condition changes captured by a newly formulated latent variable termed
“mental loading” herein, our research examines how a consumer’s men-
tal factor affects his or her purchase behavior. From the perspective of
behavioral economics, it models consumer purchase behaviors that are
seemingly irrational from a traditional economics viewpoint. The model
is derived from a threshold-based modeling framework that incorporates
consumer heterogeneity in a hierarchical Bayesian manner, and the mod-
eling parameters are estimated by using the Markov Chain Monte Carlo
method. By using scanner panel data from a retailer, the empirical results
show that our model outperforms those without consumers’ mental con-
dition changes at the time of purchase.

Keywords: Mental accounting · Inter-shopping duration

Threshold-based model

1 Introduction
In traditional economics, consumers are supposed to be rational in that they
behave to maximize their utilities. However, consumers do not necessarily act
as rational entities, and the principle of utility maximization may not always
be applicable to describe all consumer behaviors. When goods are discounted,
for example, consumers often buy them even if they are not in need. A ratio-
nal economic entity would not risk wasting goods as it does not support utility
maximization. To explain such seemingly irrational behavior, consumers’ mental
conditions can be considered. Mental accounting is a concept of behavioral eco-
nomics that illustrates purchase behavior influenced by mental factors. In this
research, we develop models of inter-shopping duration that take account of this
mental accounting effect.
c Springer Nature Singapore Pte Ltd. 2018
J. Chen et al. (Eds.): KSS 2018, CCIS 949, pp. 1–16, 2018.
https://doi.org/10.1007/978-981-13-3149-7_1
2 K. Miyatsu and T. Sato

Mental accounting was proposed by Thaler [1] as a concept of microeco-

nomics combined with cognitive psychology. In contrast to assuming a monetary
equivalency of goods as in traditional economics, mental accounting assumes
that goods have different criteria values for consumers depending on the pur-
pose of their use and circumstances at purchase. Thaler [2] recognized three
types of heterogeneity in mental accounting: transaction utility, target category,
and valuation frequency. To accommodate the heterogeneity in the models, we
assume that consumers have two sets of response rates to the parameters in the
likelihood function and that they change when a mental factor, associated with
monthly cumulative spend, exceeds the threshold. In the mental accouting stud-
ies, new behavioral findings and their validity of the seemingly irrational rules
were often presented in marketing and finacne, such as Marberis and Huang [3],
Grinblatt and Han [4], Prelec and Loewenstein [5], Langer and Weber [6], and
Shafir and Thaler [7]. But little has been studied in mental accounting to model
at individual level.
As reciprocal to duration, Ehrenberg [8] proposed modeling the number of
visits to a store within regular intervals, assuming a Poisson distribution with
a heterogeneous rate in the Gamma distribution. A number of researchers have
developed models based on Ehrenberg’s framework, such as Charfield et al. [9],
Morrison and Schmittlein [10], Sichel [11], and Gupta [12,13]. In marketing ana-
lytics, managers are most interested in knowing how media and promotion influ-
ence their customers’ inter-shopping duration. Since Cox [14] presented a pro-
portional hazard model, numerous studies have been conducted by using his
framework. For instance, Seetharaman and Chintagunta [15] proposed a discrete
hazard model that can be applied to discrete observations and Seetharaman [16]
developed an additive hazard model to improve the estimation of the covariate
effect.
We develop a hierarchical Bayesian model of inter-shopping duration, where
two separate regimes exist for each household. Allenby et al. [17] proposed a
hierarchical Bayesian model with three different regimes and applied for direct
marketing of financial services. However, their estimating a transitional point
between regimes is history data-driven, and it is not suitable for our model that
incorporates structured parameters for separate regimes. Threshold-based app-
roach is deployed in our model as is demonstrated by Ferreira [18], Geweke and
Terui [19], Chen and Lee [20]. In marketing, Terui and Danaha [21] deployed the
framework to estimate reference price regimes, and Terui and Ban [22] applied
for ad-stock effect.
The remainder of this article consists of the following sections. Section 2
defines the models, and we derive the estimation algorithm in Sect. 3. In Sect. 4,
the empirical study using a retailer’s scanner panel data is presented, and Sect. 5
concludes.
 Heterogeneous Mental Accounting Impact of Inter-shopping Durations 3

2 Models
2.1 Mental Loading Model
In this study, a new concept to capture mental pressure at the time of purchase
is introduced, called mental loading. This is defined as the cumulative spend at
a point in time during two consecutive paydays. Upon being paid every month,
mental loading is cleared and increases every time a new purchase is made. This
reflects a consumer’s mental condition at purchase; in other words, the consumer
feels more pressure spending when mental loading is high. However, information
on an individual’s payday is not available for the models and differs by individual.
Let us first define Loading_Period in Eq. (1) to specify the three most com-
mon types of paydays. First, loading_period l = 1 corresponds to a payday on the
25th of every month, which would then continue until the 24th of the following
month. In Japan, the three days in Eq. (1) are representative paydays:
⎧
⎪ th th
⎨l = 1 ⇒ 25 · Nmonth − 24 · (N + 1)month
Loading_P eriod l = 2 ⇒ 5 · Nmonth − 4th · (N + 1)month
th (1)
⎪
⎩ th th
l = 3 ⇒ 17 · Nmonth − 16 · (N + 1)month

Second, let cummi,ti ,l be denoted as the cumulative spend for household i, pur-
chase occasion at ti , and loading_period of l in Eq. (2). For the first time after
payday, cummi,ti ,l is always 0 and purchase amount Mi,j is added at every
purchase until the next payday:

transl (ti )−1
j=1 (Mi,j ) tranl (ti ) = 1
cummi,ti ,l = (2)
0 tranl (ti ) = 1

Finally, mental loading is defined as the linear combination of cumulative

∗(k)
spend cummi,ti ,l (l = 1, 2, 3) and weighting factor αi (k = 1, 2, 3) in Eq.
∗(k) ∗(k) 3 ∗(k)
(3), where αi has the constraints of 0 ≤ αi ≤ 1 and k=1 αi = 1, and
∗(1) ∗(2)
α∗i = (αi , αi ) are the mental loading structural parameters to estimate:
∗(1) ∗(2) ∗(3)
CummMi,ti = αi cummi,ti ,1 + αi cummi,ti ,2 + αi cummi,ti ,3 (3)

Each household has its own structural parameters, which reflect the composition
of the income earner(s) in a family. For example, in the case of α∗i = (1, 0), a
household has a single income earner paid on 25th of every month. When a
household’s parameter is α∗i = (0.5, 0.5), this family should have two income
earners with paydays on the 25th and 5th of every month. Figure 1 illustrates
these two cases.

2.2 Inter-shopping Duration Model

Let us define inter-shopping duration as the number of days since the imme-
diate previous occasion. For the shopping time ti,ni of household i on the nth
4 K. Miyatsu and T. Sato

single_income

40000
double_income

30000
yen

20000
10000
0

0 10 20 30 40 50

shopping occasions

Fig. 1. Mental loading: single income and double income

occasion, the inter-shopping duration is calculated as yi,ti = ti,ni − ti,ni −1 . For

household i, our model has two shopping regimes depending on the degree of
mental loading at time ti compared with threshold γcum,i . The inter-shopping
duration is assumed to be distributed log-normally for both regimes, whereas
their distribution parameters differ. When mental loading CummMi,ti at time
ti for household i exceeds threshold γcum,i , their shopping regime changes from
Regime 2 to Regime 1. Equation (4) presents the inter-shopping duration model
for household i at time ti,ti :
(1) (2) 2(1) 2(2)
P r(Yi,ti = yi,ti | μi , μi , σi , σi , γcum,i , CummMi,ti ) (4)
⎧  
⎪
⎪ 1 −1 (1) 2
⎪
⎪ exp {log(yi,ti ) − μi } , CummMi,ti ≥ γcum,i (Regime1)
⎨√ (1) 2(1)
2πσi yi,ti  2πσi 
=
⎪
⎪ 1 −1
⎪
⎪ exp
(2) 2
{log(yi,ti ) − μi } , CummMi,ti < γcum,i (Regime2)
⎩√ (2) 2(2)
2πσi yi,ti 2πσi

(k)
Let location parameter μi (k = 1, 2) be assumed to have a linear combina-
tion structure with explanatory vector xi,ti for household i at time ti ; then, the
parameters become time-dependent as the explanatory vector changes over time.
(k) (k) (k)
Now, μi (k = 1, 2) is replaced with μi,ti = xti,tt βi (k = 1, 2), and the likelihood
of Eq. (4) is multiplied for all shopping occasions until Ti , where Ti is the time of
the last occasion of household i in the period. R(k) {CummMi,ti , γcum,i } is also
introduced as an operator to determine to which regime it belongs. Assuming
that all households are independent, the total likelihood is derived in Eq. (5):
(1) (2) 2(1) 2(2)
p({yi,Ti } | {βi }, {βi }, {σi }, {σi }, {γcum,i }, {CummMi,Ti }, {xi,ti }) (5)
H 2 Ti  
   1 −1 t (k)2
= √ (k)
exp 2(k)
{log(yi,ti )−x i,tt β i }
i=1 k=1 2πσi yi,ti 2πσi
ti ∈R(k) {CummMi,t ,γinv,i }
i
 Heterogeneous Mental Accounting Impact of Inter-shopping Durations 5

2.3 Hierarchical Model

Because the model coefficients and structural parameters among house-
holds are assumed to have commonality in their attributes, a hierarchi-
cal structure is deployed as a linear function of these variables. By com-
bining the regression coefficients in Eq. (5) and the threshold for each
(1) t (2) t
household, ψi = (βi , βi , log(γcum,i )) is defined. Likewise, mental load-
(1) (2)
ing weight vector αi = (αi , αi ) is defined as an inverse logit function of
(k)
∗(k) exp(αi )
αi = 2 (l)
(k = 1, 2), and it possesses a similar hierarchical
1 + l=1 exp(αi )
structure. Equations (6) and (7) are hierarchical models. For the variance param-
2(1) 2(2)
eters σi2 = (σi , σi ), a hierarchical mechanism is not constructed as a regres-
sion model in this study (we assume the prior distribution for these parameters):

ψi = zit Δψ + ψ ψ
i , i ∼ N(0, Σψ ) (6)

αi = zit Δα + α α
i , i ∼ N(0, Σα ) (7)
ψi and αi are expressed as linear functions of household attribute vector
zi = (zi,1 , zi,2 , ..., zi,q )t and coefficient matrix Δψ , Δα , where q is a dimension
of the attribute vector. Σψ and Σα are the variance-covariance matrixes of ψi
and αi , respectively.
For prior distributions, these matrixes are assumed to have a natu-
ral conjugate relationship of p(Δψ , Σψ ) = p(Δψ |Σψ )p(Σψ ), p(Δα , Σα ) =
p(Δα |Σα )p(Σα ), where p(Δψ |Σψ ), p(Δα |Σα ) are multi-variable normal dis-
tributions and p(Σψ ), p(Σα ) are inverse-Wishart distributions. As a result, the
coefficients of the linear functions of the hierarchical structure are expressed in
the following equations:

δψ = vec(Δψ ) | Σψ ∼ N(δ¯ψ , Σψ ⊗ A−1

ψ ), Σψ ∼ IW(υψ,0 , Vψ,0 ) (8)

δα = vec(Δα ) | Σα ∼ N(δ¯α , Σα ⊗ A−1

α ), Σα ∼ IW(υα,0 , Vα,0 ) (9)

3 Algorithm of the Model Estimation

Based on the models developed in the previous section, Fig. 2 depicts the directed
acyclic graph of our proposed model. Equation (10) shows the decomposition of
the simultaneous posterior distribution of the parameters. In the derivation of
Eq. (10), we assume independence between households:
2(1) 2(2)
p({ψi }, {αi }, {σi }, {σi } | {yi,Ti }, {xi,Ti }, {cummi,Ti }, {zi })

∝ p(Δψ |Σψ )p(Σψ )p(Δα |Σα )p(Σα )×

H

 2(1) 2(2)
p(ψi |Δψ , Σψ , zi )p(αi |Δα , Σα , zi )p(σi )p(σi )×
i=1
6 K. Miyatsu and T. Sato

Fig. 2. DAG (Directed Acyclic Graph).

2 Ti  
  1 −1 (k) t (k) 2
√ (k)
exp 2(k)
{log(yi,ti ) − β i xi,t }
i
k=1 2πσi yi,ti 2πσi
ti ∈R(k) {cummi,t ,γcum,i ,αi }
i
(10)
2(1) 2(2)
The parameters included in the model are {ψi }, {αi }, {σi }, {σi },
2(1) 2(2)
Δψ , Δα , Σψ , Σα , among which {ψi }, {αi }, {σi }, {σi } are heterogeneous
to the household, while Δψ , Δα , Σψ , Σα are common in all households. The
Markov Chain Monte Carlo method is applied to estimate the parameters.
Because Δψ , Δα , Σψ , Σα are structured in natural conjugate relationships
defined in Eqs. (8) and (9), these parameters can be drawn by using Gibbs
samplers. On the contrary, {ψi }, {αi } cannot formulate the conjugate structure,
and thus it needs random walk samplers of the Metropolis-Hastings algorithm.
2(1) 2(2)
σ i , σi are sampled independently by the Gibbs samplers:

δψ | Σψ ∼ N(δ˜ψ , Σψ ⊗ (Z t Z + Aψ )−1 ) (11)

Σψ ∼ IW(υψ,0 + H, Vψ,0 + Sψ ) (12)

where δ˜ψ = vec(Δ̃ψ ), Δ̃ψ = (Z t Z + Aψ )−1 (Z t Z Δ̂ψ + Aψ Δ̄ψ ),
H
Δ̂ψ = (Z t Z)−1 Z t Ψ, Sψ = i=1 (ψi − ψ̄i )(ψi − ψ̄i )t , Ψ = (ψ1 , ψ2 , . . . , ψH ),
δ̄ψ = 0, Aψ = 10−6 I, υψ,0 = 16, Vψ,0 = υψ,0 I

δα | Σα ∼ N(δ̃α , Σα ⊗ (Z t Z + Aα )−1 ) (13)

Σα ∼ IW(υα,0 + H, Vα,0 + Sα ) (14)
where δ˜α = vec(Δ̃α ), Δ̃α = (Z t Z + Aα )−1 (Z t Z Δ̂α + Aα Δ̄α ),
H
Δ̂α = (Z t Z)−1 Z t A, Sα = i=1 (αi − ᾱi )(αi − ᾱi )t , A = (α1 , α2 , . . . , αH ),
δ̄α = 0, Aα = 10−6 I, υα,0 = 16, Vα,0 = υα,0 I
When the likelihood function is log-normal, where log(yi,ti ) is a normal dis-
tribution and the variance of the prior distribution is inverse-Gamma, the vari-
2(k)
ance of posterior distribution σi (k = 1, 2) in Eq. (10) is also full-conditionally
 Heterogeneous Mental Accounting Impact of Inter-shopping Durations 7

(k)
expressed in inverse-Gamma form with βi in Eq. (15):
2(k) (k) ∗
σi | βi ∼ IG(vi,n /2, s∗i,n /2), k = 1, 2 (15)

∗ (k) t (k)
where vi,n v0 + ni , s∗i,n = s0 + (log(yi ) − xi t βi ) (log(yi ) − xi t βi ), v0 = 100,
s0 = 100.
For the random walk sampling of ψi , αi , candidate samples are iteratively
(r−1) 2 (r−1) 2
generated by ψi = ψi + δRW ψ
I and αi = αi + δRW α
I, where the vari-
ance in random walk noise is set as σRWψ = 10 , σRWα = 10−2 , respectively.
2 −5 2

4 Empirical Analysis

4.1 Data

We estimate the models by using scanner panel data from a retailer’s store
in Tokyo between January 1 and December 31, 2001. Whenever a household
member visits the store and makes a purchase, the details of the transaction
are recorded for each household. Altogether, 100 sample panels were randomly
selected of those who shopped more than 50 occasions and at least once every
month in that period. Table 1 shows the mean and standard deviation (SD) of
the variables and household attributes.
The explanatory variables consist of two types: household-specific and com-
mon for all households. The discount rate and items on the flyer are identical for
all households at the time of shopping, whereas the number of items purchased
on the previous shopping occasion differs by household. Unlike the brand choice
model in marketing science, which concerns particular products or categories,
we consider all the products in the store. The discount rate is defined as the
average rate of products available at all times, which is 16% in the study period.
The products examined in this study only include non-perishable food and sun-
dries. Perishable food is merchandised on market quotations, and no discount
rate is available. The number of items on the flyer includes perishable food, and
the average amount is 37.2. As each household has its own family structure and
shopping pattern, the number of items purchased per store visit differs. The
average number of items per shopping is 14.3, of which 66% are non-perishable
items. Household attributes are calculated based on holdout samples before the
study period. Average spend per shopping trip is 1,537 JPY, and 61% of house-
holds are located within walking distance of the store. The region around the
retail store has 55% of part-time workers.

4.2 Model Valuation

By using scanner panel data from the 100 randomly selected household panels,
we estimated the heterogeneous parameters of the proposed model (Asymmetric
Model) together with the two other models: the model with no regime (Null
8 K. Miyatsu and T. Sato

Table 1. Definition of the variables and descriptive statistics

Variable Description Mean SD

yi,ti DRT Ni,ti Inter-shopping duration 2.02 1.98
xi,ti CN STi,ti constant:1 1.00 0.00
P U RCHi,ti The number of items purchased on the previous 14.30 8.16
occasion
IT EM Ri,ti The ratio of non-perishable items purchased on the 0.66 0.21
previous occasion
DSCN Ti,ti The average discount rate of non-perishable items 0.16 0.11
on the current occasion
IN ST Mi,ti The square root of the number of flyers distributed 6.10 4.58
on the current occasion
CummMi,ti *** Latent variable applicable only for Regime 1 *** − −
zi LP AIDi log (the average payment amount per shopping 0.43 0.17
occasion/1,000)
W ORK1i Work status type dummy 1 (home-maker = 1, 0.30 0.46
others = 0)
W ORK2i Work status type dummy 2 (part-time = 1, 0.55 0.50
others = 0)
W ORK3i Work status type dummy 3 (full-time = 1, 0.11 0.31
others = 0)
F OOTi Transportation dummy (on foot = 1, others = 0) 0.61 0.49

Table 2. Model valuation

Null model Symmetric model Asymmetric model

Deviance information 52,152.26 44,374.50 39,813.53
criteria [23] DIC
Log of the marginal −24,529.15 −22,845.47 −19,822.21
likelihood [24] LML

Model) and the threshold model with symmetric data without the mental load-
ing variable (Symmetric Model). The Markov Chain Monte Carlo method was
applied for 50,000 iterations; 45,000 samples were discarded as the burn-in period
and the remaining 5,000 samples were used in the analysis. Table 2 summarizes
the statistics for the model valuation. Both statistics support our proposed model
(Asymmetric Model) following the Symmetric Model and Null Model.

(1) (2)
4.3 Regression Parameter (βi , βi )

Table 3 shows the statistics for the average coefficient parameters of all the house-
holds in each regime. Under Bayesian statistics, heterogeneous coefficients are
 Heterogeneous Mental Accounting Impact of Inter-shopping Durations 9

estimated for each household to analyze the impact on the inter-shopping dura-
tion by household. However, in this subsection, we discuss shopping behavior
based on the average coefficients of all household panels. The intercept, which is
(1) (2)
a baseline of location parameter μ(k) , is β1 > β1 . Eliminating all the factors,
the inter-shopping duration tends to be longer in Regime 1. However, with a
(1)
negative sign of β6 , mental loading shortens this duration when it exceeds the
threshold. As a result, the inter-shopping duration tends to be shorter in Regime
(k)
1. Regarding β2 , the more items a household purchases, the longer is the dura-
tion to the next shopping occasion. When a household possesses more goods in
stock, it is rational to refrain from shopping, as there is no immediate need to
visit the store. However, this is no longer the case or a household seems to be
indifferent to the inter-shopping duration when it feels more pressure to spend.
This phenomenon is noteworthy when more non-perishable goods are procured.
(1) (2)
For the ratio of non-perishable items, β3 < β3 means that it becomes more
(1) (2)
insensitive when the regime switches to Regime 1. On the contrary, β4 < β4
implies that the discount is more sensitive in Regime 1. Lastly, the expected sign
(k)
of β5 is negative, suggesting that the duration shortens when more items are
on a flyer; however, this is only the case in Regime 2. Flyers are not as effective
in Regime 1. In general, the inter-shopping duration becomes shorter in Regime
1, where a household feels more pressure to spend. From a purchase behavior
viewpoint, households thus turn out to be more planned in Regime 1 and they
visit a store to fulfill their immediate needs at any time.

4.4 Threshold Parameter (γcu m ,i )

Table 3 shows the statistics of the average threshold parameters of all the house-
holds. The average threshold is 11,649 JPY, indicating that regime switching
occurs when mental loading exceeds 27.3% of monthly expenditure on average.
(1)
Understanding the relationship between γcum,i and β6,i would help explain
how mental loading affects the duration. Figure 3 shows the scatter plots of γcum,i
(1)
and β6,i . The higher the threshold, the less effective is the response of mental
loading. Households with a lower threshold perceive pressure to spend at an
earlier stage, raising the impact of mental loading. As a result, the inter-shopping
duration shortens and thus a household purchases goods more frequently to fulfill
their immediate needs.

(2)
4.5 Mental Loading Structural Parameter (αi (1) , αi )

Mental loading is the latent variable, and the structural parameters determine
the composition of income earners based on their monthly payday dates. Figure 4
∗(1) ∗(2)
shows the scatter plots of (αi , αi ). The plots on (1, 0), (0, 1), (0, 0) corre-
spond to single-income households with a loading period l = 1, 2, 3. The plots on
∗(1) ∗(2) ∗(2) ∗(1)
the αi + αi = 1, αi , and αi axes indicate double-income households
with the loading periods l = (1, 2), (2, 3), (3, 1), respectively; the others are from
10 K. Miyatsu and T. Sato

Table 3. Basic statistics of the estimated parameters

Coefficient Mean Median Min Max SD Rate∗

(1)
β1,i 0.4052 0.3900 −0.7397 1.4549 0.3760 82%
(1)
β2,i −0.0068 0.0036 −0.5043 0.1586 0.0789 22%
(1)
β3,i 0.1499 0.1351 −0.8320 0.6743 0.2381 65%
(1)
β4,i 0.0577 0.0509 −0.7608 1.4232 0.4127 70%
(1)
β5,i 0.0142 0.0007 −0.2914 0.9730 0.1101 19%
(1)
β6,i −0.0453 −0.0598 −0.3933 0.6258 0.1554 66%
(2)
β1,i 0.3724 0.3298 −0.6494 1.5750 0.4618 84%
(2)
β2,i 0.0502 0.0248 −0.3706 0.8450 0.1500 39%
(2)
β3,i 0.2434 0.2416 −1.2171 1.1805 0.4380 76%
(2)
β4,i 0.1151 0.1983 −0.9531 1.0491 0.4491 81%
(2)
β5,i −0.0166 0.0012 −0.5850 0.5322 0.1557 31%
2(1)
σi 0.5962 0.5631 0.2455 1.0129 0.2160 81%
2(2)
σi 0.9263 0.9649 0.5981 1.0231 0.0949 10%
γcum,i 11,649 7,554 440 111,679 14,887 100%
(1)
αi 0.3333 0.2600 0.0000 1.0000 0.3671 63%
(2)
αi 0.3341 0.3393 0.0000 1.0000 0.3464 67%
Rate∗ : 95% statistical significance of all households based on the quasi
t-value (Mean/SD)
0.6
0.4
0.2
beta6

0.0
−0.2
−0.4

0 20000 40000 60000 80000 100000

gamma

Fig. 3. Scatter plots (γcum − β6 )

triple- income households. Figure 4 also summarizes the occupancy rate of the
composition of income earners among all household panels, counted based on
estimating α∗i by disregarding values below 0.01. The shares of single-, double-,
and triple- income households are 39%, 42%, and 19%, respectively. In most com-
panies in Japan, employees are paid every month; however, the payday date dif-
fers largely depending on the size and ownership status of the company: Table 3
 Heterogeneous Mental Accounting Impact of Inter-shopping Durations 11

shows the statistics of the average threshold parameters of all the households.
The average threshold is 11,649 JPY, indicating that regime switching occurs
when mental loading exceeds 27.3% of monthly expenditure on average.
(1)
Understanding the relationship between γcum,i and β6,i would help explain
how mental loading affects the duration. Figure 3 shows the scatter plots of γcum,i
(1)
and β6,i . The higher the threshold, the less effective is the response of mental
loading. Households with a lower threshold perceive pressure to spend at an ear-
lier stage, raising the impact of mental loading. As a result, the inter-shopping
duration shortens and thus a household purchases goods more frequently to fulfill
their immediate needs. l = 1 is applicable for large publicly listed corporations,
l = 2 for private and small or medium-sized companies, and l = 3 for government
and public service entities. At a glance, the proportion of triple-income house-
holds seems higher. As family size has deceased in Japan, the triple- income
earners in a family are higher than expected. However, most part-timers have
two paydays a month, so this is treated as different income earners in this esti-
mation.

Fig. 4. Mental loading structural vector

4.6 Hierarchical Model Parameter (δ1 , δ2 , δ3 , δ4 , δ5 )

The hierarchical model parameters reveal the common impact on the different
coefficients. Table 4 shows the estimated results for the hierarchical model param-
eters. No obvious commonalities are identified in most cases, which is partly
because of the insufficient number of household panels in the model, except for
γcum . The threshold parameter is estimated in a stable manner, and the levels
of significance of the attribute variables are high except for F OOT . The degree
(1) (2)
of average spend LP AID influences β1 and β1 . This average spend has a
negative impact on the intercept in Regime 1, while it is positive in Regime 2.
(2) (2)
Part-time workers have a positive impact on β3 and β4 only in Regime 2.
12 K. Miyatsu and T. Sato

Table 4. Hierarchical model estimation

VariableĄ@ CoefficientĄ@ LP AID δ1 W ORK1 δ2 W ORK2 δ3 W ORK3 δ4 F OOT δ5

(1)
CN ST (1) β1 ∗ − 0.3789 ∗∗∗ 0.4862 ∗∗∗ 0.5501 ∗∗∗ 0.5762 0.0630
(2)
CN ST (2) β1 ∗ 0.4888 0.2409 0.0841 −0.0586 0.0432
(1) (1)
P U RCH β2 −0.0177 0.0071 −0.0101 0.0127 0.0038
(2)
P U RCH (2) β2 −0.0229 −0.0210 0.0232 −0.0229 0.0793
(1)
IT EM R(1) β3 0.1583 0.1222 −0.0176 0.0267 0.0859
(2) (2) ∗∗ 0.3939
IT EM R β3 0.1366 0.0840 0.1425 0.0487
(1) (1) ∗∗ 0.1530
DSCN T β4 −0.2884 0.0893 0.0677 0.0996
(2) (2) ∗∗∗ 0.5140
DSCN T β4 −0.2109 0.1492 0.1219 0.0617
(1)
IN ST M (1) β5 0.0202 0.0160 0.0294 0.0086 −0.0266
(2)
IN ST M (2) β5 −0.0052 0.0651 −0.0035 0.0372 −0.0535
(1)
CummM (1) β6 ∗ 0.3398 ∗ −0.1828 ∗ −0.1709 ∗ −0.2140 −0.0261
T HRSH γcum ∗∗∗ 6.5761 ∗∗∗ 6.0261 ∗∗∗ 5.8488 ∗∗∗ 5.9571 0.4514
ALP H (1) α(1) ∗∗ 4.6526 ∗∗ − 36.8833 −17.3374 ∗∗ − 43.1437 −4.6015
ALP H (2) α(2) 18.8939 −20.6767 −16.7853 −19.8601 1.9962
Statistical significance: ∗∗∗ 95%, ∗∗ 90%, ∗ 80%.

regime_1 regime_2
1.0

1.0

usr1 usr1
usr2 usr2
usr3 usr3
0.8

0.8
0.6

0.6
hazard

hazard
0.4

0.4
0.2

0.2
0.0

0.0

0 5 10 15 20 25 0 5 10 15 20 25

duration (days) duration (days)

Fig. 5. Hazard (a Double-Regime Occupant)

4.7 Inter-shopping Duration Hazard

The parametric hazard function is derived based on the distribution parame-

ters of the log-normal probability density function. Although variance is a fixed
parameter for each household, the location parameter is designed to have a
linear structure with the explanatory variables for household i at time ti . Con-
sequently, the location parameter possesses time dependency, meaning that the
hazard function also changes on every occasion. In this subsection, we discuss
the hazard functions for the households listed in Table 5, where the location
 Heterogeneous Mental Accounting Impact of Inter-shopping Durations 13

Table 5. Regime occupancy and durations

Location Variance Occupancy Duration Threshold

R1 R2 R1 R2 R1 R2 R1 R2 γcum,i
usr1 0.2525 2.1217 0.4109 1.0128 99% 1% 1.44 8.00 3,744
usr2 0.4178 1.3518 0.5095 0.9744 93% 7% 1.71 5.62 18,056
usr3 0.7312 1.0701 0.5095 0.8622 26% 74% 2.34 3.69 14,056
usr4 0.1038 – 0.3005 – 100% 0% 1.17 – 3,269
usr5 0.3885 – 0.4113 – 100% 0% 1.67 – 4,985
usr6 0.3584 – 0.4335 – 100% 0% 1.59 – 8,366
usr7 – 0.7772 – 0.6464 0% 100% – 2.62 3,744
usr8 – 0.8376 – 0.8382 0% 100% – 3.34 18,056
usr9 – 1.0801 – 0.6565 0% 100% – 3.39 14,810
R1: CummMi,ti ≥ γcum,i (Regime1), R2: CummMi,ti < γcum,i (Regime2)
 i (k) Ni (k)
Location : (1/Ni ) N
n=1 μi,n , Duration : (1/Ni ) n=1 yi,n ,(k = 1, 2)

regime_1 regime_2
1.0

1.0

usr4 usr7
usr5 usr7
usr6 usr9
0.8

0.8
0.6

0.6
hazard

hazard
0.4

0.4
0.2

0.2
0.0

0.0

0 5 10 15 20 25 0 5 10 15 20 25

duration (days) duration (days)

Fig. 6. Hazard (a Single-Regime Occupant)

parameters are calculated by using the average values of the explanatory vector
and duration is the average inter-shopping duration for each regime.
As usr1–3 have shopping occasions in each regime, they have different hazard
functions, which are not monotonously increasing or decreasing but unimodal.
The location and height of a peak are different for each household. Figure 5
illustrates the hazard functions of the users. They all have sharp-pointed peaks
in Regime 1. Their purchase behavior changes once they enter Regime 1, and
they shift the mindset of paying attention to fulfill their immediate needs; thus,
the inter-shopping duration becomes shorter. Usr4–6 and usr7–9 have shopping
occasions in a single regime only. Figure 6 show the hazard functions of usr4–6 on
the left and usr7–9 on the right. Whenever a single regime remains, their hazard
Another random document with
no related content on Scribd:
How the Lord Cromwell[2006] exalted
from meane estate, was after by the
enuie of the Bishop of Winchester,
and other his complices, brought to
vntimely end, Anno Dom. 1540.
1.

“Wak’d, and trembling betwixt rage and dread

With the loud slander (by the impious time)
That of my actions euery where is spread,
Through which to honor falsely I should clime,
From the sad dwelling of th’vntimely dead,
To quit me of that execrable crime,
Cromwell appeares his wretched plight to show,
Much that can tell, one much that once did know.[2007]

2.

Roughly not made vp in the common mould,

That with the vulgar vilely I should die,
What thing so strange of Cromwell is not told?
What man more prais’d? who more condemn’d then I?
That with the world when I am waxed old,
Most t’were vnfit that fame of me should lie
With fables vaine my historie to fill,
Forcing my good, excusing of my ill.

3.
You, that but hearing of my hated name,
Your ancient malice instantly bewray,
And for my sake your ill deserued blame
Vpon my legend publikely shall lay:
Would you forbeare to blast me with defame,
Might I so meane a priuiledge but pray,
He that three ages hath endur’d your wrong,
Heare him a little that hath heard you long.

4.

Since Rome’s sad ruine heere by me began,

Who her religion pluckt vp by the root,
Of the false world such hate for which I wan,
Which still at me her poisned’st darts doth shoot:
That to excuse it, do the best I can,
Little, I feare, my labour me will boot:
Yet will I speake my troubled heart to ease,
Much to the mind, her selfe it is to please.

5.

O powerfull number, from whose stricter law

Heart-mouing musicke did receiue the ground
Which men to faire ciuilitie did draw
With the brute beast when lawlesse he was found:
O, if according to the wiser saw
There be a high diuinitie in sound,
Be now abundant prosp’rously to aide
The pen prepar’d my doubtfull case to pleade.

6.

Putney the place made blessed in my birth,

Whose meanest cottage simplie me did shrowd,
To me as dearest of the English earth:
So of my bringing that poore village prou’d,
Though in a time when neuer lesse the dearth
Of happie wits, yet mine so well allow’d
That with the best she boldly durst confer
Him that his breath[2008] acknowledged from her.

7.

Twice flow’d proud Thames, as at my comming wood,

Striking the wondring borderers with feare,
And the pale genius of that aged flood
Vnto my mother[2009] labouring did appeare,
And with a countenance much distracted stood,
Threatning the fruit her pained wombe should beare:
My speedie birth being added thereunto,
Seem’d to foretell that much I came to do.

8.

That[2010] was reserued for those worser daies,

As the great ebbe vnto so long a flow,
When what those ages formerly did raise,
This, when I liu’d, did lastly ouerthrow,
And that great’st labour of the world did seaze,
Only for which immedicable blow
Due to that time, me dooming heauen ordain’d,
Wherein confusion absolutely raign’d.

9.

Vainly yet noted this prodigious signe,

Often predictions of most fearefull things,
As plagues, or warre, or great men to decline,
Rising of commons, or the death of kings:
But some strange newes though euer it diuine,
Yet forth them not immediatly it brings,
Vntill th’effects men afterward did learne,
To know that me it chiefly did concerne.
 10.

Whil’st yet my father by his painfull trade,

Whose laboured anuile only was his fee,
Whom my great towardnesse strongly did perswade
In knowledge to haue educated mee:
But death did him vnluckily inuade,
Ere he the fruits of his desire could see,
Leauing me yong, then little that did know
How me the heauens had purpos’d to bestow.

11.

Hopelesse as helpelesse most might me suppose,

Whose meannesse seem’d their abiect breath to draw:
Yet did my breast that glorious fire inclose,
Which their dull purblind ignorance not saw,
Which still is setled vpon outward showes,
The vulgar’s iudgement euer is so raw,
Which the vnworthiest sottishly do loue
In their owne region properly that moue.

12.

Yet me my fortune so could not disguise,

But through this cloud were some that did me know,
Which then the rest more happie or more wise,
Me did relieue when I was driuen low,
Which, as the staier, by which I first did rise,
When to my height I afterward did grow,
Them to requite my bounties were so hie,
As made my fame through euery eare to flie.

13.

That height and godlike puritie of minde

Resteth not still, where titles most adorne
With any, nor peculiarly confinde
To names, and to be limited doth scorne:
Man doth the most degenerate from kinde,
Richest and poorest both alike are borne:
And to be alwaies pertinently good,
Followes not still the greatnes of our blood.

14.

Pitie it is, that to one vertuous man

That marke him lent to gentrie to aduance,
Which first by noble industrie he wan,
His baser issue after should inhance,
And the rude slaue not any good that can,
Such should thrust downe by what is his by chance:
As had not he been first that him did raise,
Nere had his great heire wrought his grandsire’s
praise.

15.

How weake art thou that makest it thy end

To heape such worldly dignities on thee,
When vpon fortune only they depend,
And by her changes gouerned must bee?
Besides the dangers still that such attend,
Liuel’est of all men purtraied out in mee,
When that, for which I hated was of all,
Soon’st from me fled, scarse tarrying for my fall.

16.

You that but boast your ancestors proud stile,

And the large stem whence your vaine greatnes grew,
When you your selues are ignorant and vile,
Nor glorious thing dare actually pursue,
That all good spirits would vtterly exile,
Doubting their worth should else discouer you,
Giuing your selues vnto ignoble things:
 Base I proclaime you though deriu’d from kings.

17.

Vertue, but poore, God in this earth doth place

’Gainst the rude world to stand vp in his right,
To suffer sad affliction and disgrace,
Not ceasing to pursue her with despight:
Yet when of all she is accounted base,
And seeming in most miserable plight,
Out of her power new life to her doth take,
Least then dismai’d when all do her forsake.

18.

That is the man of an vndaunted spirit,

For her deare sake that offereth him to dye,
For whom, when him the world doth disinherit,
Looketh vpon it with a pleased eye,
What’s done for vertue thinking it doth merit,
Daring the proudest menaces defie,
More worth then life, how ere the base world rate him,
Belou’d of heauen, although the earth doth hate him.

19.

Iniurious time, vnto the good vniust,

O, how may weake posteritie suppose
Euer to haue their merit from the dust,
’Gainst them thy partialitie that knowes!
To thy report, O, who shall euer trust,
Triumphant arches building vnto those
Allow’d the longest memorie to haue,
That were the most vnworthie of a graue?

20.

But my cleere mettle had that powerfull heat,

As it not turn’d with all that fortune could:
Nor when the world me terriblest did threat,
Could that place win[2011] which my hie thoughts did
hold,
That waxed still more prosperously great,
The more the world me stroue to haue control’d,
On my owne columnes constantly to stand,
Without the false helpe of another’s hand.

21.

My youthfull course thus wisely did I steere,

T’auoid those rockes my wracke that else did thret:
Yet some faire hopes from farre did still appeere,
If that too much my wants me did not let:
Wherefore my selfe aboue my selfe to beare,
Still as I grew, I knowledge stroue to get,
To perfect that which in the embryon was,
Whose birth, I found, time well might bring to passe.

22.

But when my meanes to faile me I did finde,

My selfe to trauell presently betooke,[2012]
As much distastfull[2013] to my noble minde,
That the vile world into my wants should looke,
And of my selfe industriously[2014] inclinde,
To measure other’s actions with my booke,
I might my iudgement rectifie[2015] thereby,
In matters that were difficult and hie.

23.

When, loe, it hapt that fortune, as my guide,

Of me did with such prouidence dispose,
That th’English merchants then, who did reside
At Antwerpe, me their secretarie chose,
(As though in me to manifest her pride)
Whence to those principalities I rose,
To pluck me downe, whence afterward she fear’d
Beyond her power that almost she had rear’d.

24.

When first the wealthie Netherlands me traind

In wise commerce, most proper to the place,
And from my countrie carefully me wain’d,
That with the world did chiefly winne[2016] me grace,
Where great experience happily I gaind:
Yet here I seem’d but tutor’d for a space,
For hie imploiment otherwise ordaind,
Till which the time I idely entertaind.

25.

For hauing Boston businesse in[2017] hand,

The charge thereof on Chambers being laid,
Coming to Flanders, hapt to vnderstand
Of me, whom he requested him to aid:
Of which, when I the benefit had scand,
Weighing what time at Antwerpe I had staid,
Quickly me wonne[2018] faire Italy to trie,
Vnder a cheerefull and more luckie skie:

26.

For what the meanest cleerely makes to shine,

Youth, wit, and courage, all in me concurre
In euery proiect, that so powerfull trine
By whose kind working brauely I did sturre,
Which to each hie and glorious designe
(The time could offer) freely did me spurre,
As forcing fate some new thing to prepare
(Shewing successe) t’attempt that could me dare.
 27.

Where now my spirit got roomth it selfe to show,

To the fair’st pitch doth make a gallant flight,
From things that too much earthly were and low,
Strongly attracted by a genuine light,
Where higher still it euery day did grow:
And being in so excellent a plight,
Crau’d but occasion happily to proue
How much it sate each vulgar spirit aboue.

28.

The good successe th’affaires of England found,

Much prais’d the choice of me that had been made:
For where most men the depth durst hardly sound,
I held it nothing boldly through to wade
My selfe, and through the strait’st waies I woond:
So could I act, so well I could perswade,
As meerely iouiall, me to mirth applie,[2019]
Compos’d of freedome and alacritie.

29.

Not long it was ere Rome of me did ring

(Hardly shall Rome so full daies see again)
Of freemen’s catches to the pope I sing,
Which wan much licence to my countrimen,
Thither the which I was the first did bring,
That were vnknowne to Italy till then:
Light humours, them when iudgement doth direct,
Euen of the wise win plausible respect.

30.

And those, from whom that pensions were allow’d,

And heere[2020] did for intelligence remaine,
Vnder my power themselues were glad to shroud,
Russell and Pace, yea, oftentimes were faine,
When as their names they durst not haue auow’d,
Me into their societie t’retaine,
Rising before me, mightie as they were,
Great though at home, yet did they need me there.

31.

In forraine parts nere friends I yet forsake,

That had before been deeply bound to mee,
And would againe I vse of them should make,
But still my starres command I should be free,
And all those offers lightly from me shake,
Which to requite, I fettred else might bee,
And though that oft great perils me oppungne,
And meanes were weak, my mind was euer strong.

32.

And[2021] those great wants fate to my youth did tie

Me from delights[2022] of those rich countries driue,
Thereby inforc’d with painfull industrie
Against affliction manfully to striue,
Vnder her burthen faintly not to lie,
But since my good I hardly must deriue,
Vnto the same to make my selfe a way[2023]
Through all the power against me she could lay.

33.

As a comedian where my[2024] life I led,

For so a while my need did me constraine,
With other my poore countrimen (that plai’d)
Thither that came in hope of better gaine,
Whereas when fortune seem’d me low to tread
Vnder her feet, she set me vp againe,
Vntill the[2025] vse me bad her not to feare
 Her good and ill that patiently could beare.

34.

Till Charles the fift th’emperiall power did bend

’Gainst Rome, which Burbon skilfully did guide,
Which sore declining[2026] Italy did rend:
For th’right that him her holinesse denide,
Wholly her selfe enforced to defend
Gainst him that iustly punished her pride,
To which my selfe I lastly did betake,
Seeing[2027] thereof what fortune ment to make.

35.

And at the siege with that great generall seru’d,

When he did[2028] girt her stubborne waste with steele,
Within her walles who well neer being staru’d,
And that with faintnes she began to reele,
Shewing her selfe a little as she swaru’d:
First her then noting I began to feele,
She whose great power so far abroad did rome,
What in her selfe she truly was at home.

36.

That the great schoole of the false world was then,

Where her’s their subtill practises did vie,
Amongst that mightie confluence of men,
French plots propt vp by English policie,
The German powers, false shuffling, and agen
All countermin’d by skilfull Italy,
Each one in possibility to win,
Great rests were vp and mightie hands were in.

37.

Here first to worke my busie braine was set,

(My inclination finding it to please
This stirring world which strongly still did whet)
To temper in so dangerous assaies,
Which did strange formes of policies beget:
Besides in times so turbulent as these,
Wherein my studies hopefully did[2029] bend
Vnto that point the wisest[2030] made their end.

38.

And my experience happily me taught

Into the secrets of those times to see,
From whence to England afterward I brought
Those slights of state deliu’red vnto[2031] me,
In t’which were then but very[2032] few that sought,
Nor did with th’umour of that age agree,
After did great and fearfull[2033] things effect,
Whose secret working few did then suspect.

39.

When though t’were long it hapned yet at last

Some hopes me homeward secretly allur’d,
When many perils strangely I had past,
As many sad calamities endur’d
Beyond the moone, when I began to cast
By my rare parts what place might be procur’d,
If they at home were to the mightie knowne,
How they would seeme compared with their owne.

40.

Or if that there the great should me neglect,

As I the worst that vainely did not feare,
To my experience how to gaine respect
In other countries that doe hold it deare,
And now occasion seemed to reiect,[2034]
Whil’st still before me other rising were,
And some themselues had mounted to the skie,
Little before vnlike to thriue as I.

41.

When now in England bigamie with blood

Lately begot by luxurie and pride,
In their great’st fulnes peremptorie stood:
Some thereunto that diligently pri’d,[2035]
Stillie[2036] were fishing in that troubled flood
For future changes wisely to prouide,
Finding the world so rankly then to swell,
That till it brake it neuer could be well.

42.

But floting long vpon my first arriue,

Whil’st many doubts me seemed to appall,
Like to a barke that with the tide doth driue,
Hauing not[2037] left to fasten it withall,
Thus with the time by suffring I doe striue
Vnto[2038] that harbor doubtfull yet to fall:
Vntill inforc’d to put it to the chance,
Casting the fair’st my fortune to aduance.

43.

Making my selfe to mightie Wolsey knowne,

That Atlas, which the gouernement vpstai’d,
Which[2039] from meane place in little time was growne
Vp vnto him, that[2040] weight vpon him lai’d,
And being got the neerest to his throne,
He the more easly the[2041] great kingdome swai’d,
Leaning thereon his wearied selfe to breath,
Whil’st euen the greatest farre sat him[2042] beneath.
 44.

Where learned More and Gardiner I met,

Men in those times immatchable for wit,
Able that were the dullest spirit to whet,
And did my humour excellently fit,
Into their ranke that worthily did get
There as their proud competitor to sit,
One excellence to many is the mother,
Wit doth,[2043] as creatures, one beget another.

45.

This founder of the palaces of kings,

Whose veines with more then vsuall spirit were fild,
A man ordained to the mighti’st things,
In Oxford then determining to build
To Christ a colledge, and together brings,
All that thereof the great foundation wills,
There me imploies, whose industrie he found
Worthie to worke vpon the noblest ground.

46.

Yet in the entrance wisely that did feare

Coyne might fall short, yet with this worke on fire,
Wherefore such houses as religious were
Whose being no necessitie require,
But that the greater very well might beare,
From Rome the Card’nall cunningly did hire,
Winning withall his soueraigne to consent,
Both colouring with so holy an intent.

47.

This like a symptome to a long disease

Was the forerunner to this mightie fall,
And but too vnaduisedly did sease
Vpon the part that ruinated all,
Which, had the worke been of so many daies,
And more againe, recouer hardly shall:
But loe, it sunke, which time did long vphold,
Where now it lies euen leueld with the mould.

48.

Thus thou, great Rome, here first wast ouerthrowne

Thy future harmes that blindly couldst not se,
And in this worke they only were thine owne,
Whose knowledge lent that deadly wound to thee,
Which to the world before had they not showne,
Nere had those secrets been descri’d by mee,
Nor by thy wealth so many from the plow
Worne those hie types wherein they florish now.

49.

After when as the cardinall againe

Into hie fauour[2044] with the king mee brought,
With[2045] whom my selfe so well I did demeane,
As that I seem’d to exercise his thought,
And his great liking strongly did retaine
With what before my master me had[2046] taught,
From whose example, by those cels were small,
Sprang the subuersion lastly of them all.

50.

Yet many a let was cast into the way,

Wherein I ran so steddily and right,
And many a snare my aduersaries lay,
Much wrought they with their power, much with their
slight,
Wisely perceiuing that my smallest stay
Fully requir’d the vtmost of their might,
 To my ascendant hasting me[2047] to clime,
There as the first predomining the time.

51.

Knowing what wealth me earnestly did wooe,

Which I through Wolsey hapned had to finde,
And could the path most perfectly vntoo,
The king thereafter earnestly inclin’d,
Seeing besides what after I might doe
If so great power mee fully were assign’d,
By all their meanes against me strongly wrought,
Lab’ring as fast to bring their church to nought.

52.

Whil’st to the king continually I sue,

And in this businesse faithfully did stirre
Strongly t’approue[2048] my iudgement to be true
Gainst those who most supposed me to erre,
Nor the least meanes which any way I knew
Might grace me, or my purposes preferre
Did I omit, till wonne I had[2049] his eare,
Most that me mark’d, when least he seem’d to heare.

53.

This wound to them thus violently giuen,

Enuie at me her sharpest darts doth[2050] roue,
Affecting the supremacie of heauen,
As the first giants warring against Ioue,
Heap’d hils on hils, the gods till they had driuen
The meanest shapes of earthly things to proue:
So must I shift from them against me rose,
Mortall their hate, as mightie were my foes.

54.
But their great force against me wholly bent
Preuail’d vpon my purposes so farre,
That I my ruine scarsely could preuent,
So momentarie worldly fauours are,
That till the vtmost of their spight was spent,
Had not my spirit maintain’d a manly warre,
Risen they had when laid I had been low,[2051]
Vpon whose ruine after I did grow.

55.

When the great king their strange reports that tooke

That as[2052] pernitious as they potent were,
Which[2053] at the faire growth of my fortune strooke,
Whose deadly malice blame me not to feare,
Me at the first so violently shooke,
That they this frame were likely downe to beare,
If resolution with a setled brow
Had not vpheld my peremptorie vow.

56.

Yet these encounters thrust me not awry,

Nor could my courses force me to forsake,
After this shipwrack I againe must trie,
Some happier voiage hopefull still to make,
The plots that barren long we see did lie,
Some fitting season plentifully take,
One fruitfull haruest frankly doth restore
What many winters hindred had before.

57.

That to account I strictly call my wit

How it this while had managed my state,
My soule in counsell summoning to sit,
If possible to turne the course of fate,
For waies there be the greatest things to hit,
If men could find the peremptorie gate,
And since I once was got so neere the brinke
More then before, ’twould grieue me now to sinke.

58.

Bedford,[2054] whose life (some said) that I had sau’d

In Italy, one me that[2055] sauoured most,
And reuerend Hayles, who but occasion crau’d
To shew his loue, no lesse that I had cost,
Who to the king perceiuing me disgrac’d,
Whose fauour I vnluckily had lost,
Both with him great, a foot set in withall
If not to stay, to qualifie my fall.

59.

High their regard, yet higher was their hap,

Well neere quite sunke, recouer me that could,
And once more get me into fortune’s lap,
Which well my selfe might teach me there to hold,
Escap’d out of so dangerous a trap,
Whose praise by me to ages shall be told,
As the two props by which I only rose,
When most supprest, most trod on by my foes.

60.

This me to vrge the premunire wonne,

Ordain’d in matters dangerous and hie,
In t’which the heedlesse prelacie were runne,
That backe vnto the papacie did flie,
Sworne to that sea, and what before was done
Due to the king, dispensed were thereby,
In t’which first entring offred me the meane
That to throw downe, alreadie that did leane.
 61.

This was to me that ouerflowing sourse,

From whence his bounties plentifully spring,
Whose speedie current with vnusuall force
Bare me into the bosome of the king,
By putting him into that readie course
Which soone to passe his purposes might bring,
Where those which late emperiously control’d me
Pale strooke[2056] with feare stood trembling to behold
me.

62.

When state to me those ceremonies show’d

That to so great a fauorite were due,
And fortune still with honors did me load,
As though no meane she in my rising knew,
Or heauen to me more then to man had ow’d,
(What to the world vnheard of was and new)
And was to other sparing of her store
Till she could giue, or aske I could[2057] no more.

63.

Those high preferments he vpon me laid,

Might make the world me publikely to know
Such as in[2058] iudgement rightly being wai’d,
Seemed too great for me to vndergo,
Nor could his hand from powring on be stai’d,
Vntill I so abundantly did flow,
That looking downe whence lately I was cloame,
Danger bid[2059] feare, if further I should roame.

64.

For first from knighthood rising in degree,

The office of the iewell house my lot,
After the Roles he frankely gaue to mee,
From whence a priuie counsellor I got,
Chose of[2060] the garter: and the[2061] earle to bee
Of Essex: yet sufficient these not[2062]
But to the great vicegerencie I grew,
Being a title as supreame as new.

65.

So well did me these dignities befit,

And honor so me euery way became,
As more then man I had been made for it,
Or as from me it had deriu’d the name:
Where was that man[2063] whose loue I not requit
Beyond his owne imaginarie aime,
Which had me succour’d, neerely being driuen,
As things to me that idlely were not giuen?

66.

What tongue so slow the tale shall not report

Of hospitable Friscobald and mee,
And shew in how reciprocall a sort
My thankes did with his courtesie agree,
When as my meanes in Italy were short
That me relieu’d, lesse great that[2064] would not bee,
When I of England chancellor was made,[2065]
His former bounties librally repai’d?

67.

The maner briefly gentler muse relate,

Since oft before it wisely hath been told,
The sudden change of vnauoided fate,
That famous merchant, reuerend Friscobald,
Grew poore, and the small remnant of his state