Professional Documents
Culture Documents
Textbook Artificial Intelligence in Renewable Energetic Systems Mustapha Hatti Ebook All Chapter PDF
Textbook Artificial Intelligence in Renewable Energetic Systems Mustapha Hatti Ebook All Chapter PDF
https://textbookfull.com/product/artificial-intelligence-and-
algorithms-in-intelligent-systems-radek-silhavy/
https://textbookfull.com/product/progresses-in-artificial-
intelligence-and-neural-systems-anna-esposito/
https://textbookfull.com/product/knowledge-based-expert-systems-
in-chemistry-artificial-intelligence-in-decision-making-judson/
https://textbookfull.com/product/advances-in-artificial-
intelligence-software-and-systems-engineering-tareq-z-ahram/
Artificial Intelligence and Industrial Applications:
Artificial Intelligence Techniques for Cyber-Physical,
Digital Twin Systems and Engineering Applications
Tawfik Masrour
https://textbookfull.com/product/artificial-intelligence-and-
industrial-applications-artificial-intelligence-techniques-for-
cyber-physical-digital-twin-systems-and-engineering-applications-
tawfik-masrour/
https://textbookfull.com/product/artificial-intelligence-in-
value-creation-andrzej-wodecki/
https://textbookfull.com/product/artificial-intelligence-by-
example-develop-machine-intelligence-from-scratch-using-real-
artificial-intelligence-use-cases-denis-rothman/
https://textbookfull.com/product/analytics-data-science-
artificial-intelligence-systems-for-decision-support-eleventh-
edition-dursun-delen/
Lecture Notes in Networks and Systems 35
Artificial
Intelligence
in Renewable
Energetic Systems
Smart Sustainable Energy Systems
Lecture Notes in Networks and Systems
Volume 35
Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: kacprzyk@ibspan.waw.pl
The series “Lecture Notes in Networks and Systems” publishes the latest
developments in Networks and Systems—quickly, informally and with high quality.
Original research reported in proceedings and post-proceedings represents the core
of LNNS.
Volumes published in LNNS embrace all aspects and subfields of, as well as
new challenges in, Networks and Systems.
The series contains proceedings and edited volumes in systems and networks,
spanning the areas of Cyber-Physical Systems, Autonomous Systems, Sensor
Networks, Control Systems, Energy Systems, Automotive Systems, Biological
Systems, Vehicular Networking and Connected Vehicles, Aerospace Systems,
Automation, Manufacturing, Smart Grids, Nonlinear Systems, Power Systems,
Robotics, Social Systems, Economic Systems and other. Of particular value to both
the contributors and the readership are the short publication timeframe and the
world-wide distribution and exposure which enable both a wide and rapid
dissemination of research output.
The series covers the theory, applications, and perspectives on the state of the art
and future developments relevant to systems and networks, decision making, control,
complex processes and related areas, as embedded in the fields of interdisciplinary
and applied sciences, engineering, computer science, physics, economics, social, and
life sciences, as well as the paradigms and methodologies behind them.
Advisory Board
Fernando Gomide, Department of Computer Engineering and Automation—DCA, School of
Electrical and Computer Engineering—FEEC, University of Campinas—UNICAMP,
São Paulo, Brazil
e-mail: gomide@dca.fee.unicamp.br
Okyay Kaynak, Department of Electrical and Electronic Engineering, Bogazici University,
Istanbul, Turkey
e-mail: okyay.kaynak@boun.edu.tr
Derong Liu, Department of Electrical and Computer Engineering, University of Illinois
at Chicago, Chicago, USA and Institute of Automation, Chinese Academy of Sciences,
Beijing, China
e-mail: derong@uic.edu
Witold Pedrycz, Department of Electrical and Computer Engineering, University of Alberta,
Alberta, Canada and Systems Research Institute, Polish Academy of Sciences, Warsaw,
Poland
e-mail: wpedrycz@ualberta.ca
Marios M. Polycarpou, KIOS Research Center for Intelligent Systems and Networks,
Department of Electrical and Computer Engineering, University of Cyprus, Nicosia, Cyprus
e-mail: mpolycar@ucy.ac.cy
Imre J. Rudas, Óbuda University, Budapest Hungary
e-mail: rudas@uni-obuda.hu
Jun Wang, Department of Computer Science, City University of Hong Kong
Kowloon, Hong Kong
e-mail: jwang.cs@cityu.edu.hk
Artificial Intelligence
in Renewable Energetic
Systems
Smart Sustainable Energy Systems
123
Editor
Mustapha Hatti
EPST-CDER
UDES, Unité de Développement des
Equipements Solaires
Bou Ismaïl
Algeria
The development of renewable energy at low cost must necessarily involve the
intelligent optimization of energy flows and the intelligent balancing of production,
consumption, and energy storage. Intelligence is distributed at all levels and allows
information to be processed to optimize energy flows according to constraints. It is
around this thematic that is shaping the outlines of the economy of the future, and
behind these topics is the possibility of transforming society. Taking advantage
of the growing power of the microprocessor makes the complexity of renewable
energy systems accessible. Especially, since the algorithms of artificial intelligence
are gourmand in data making, it possible to take relevant decisions or even reveal
unsuspected trends in the management and optimization of renewable energy flows.
This book is a good prospect for those who work on energy systems and those who
deal with models of artificial intelligence to combine their knowledge and their
intellectual potential for the benefit of the scientific community and humanity.
Short Biography
v
Contents
vii
viii Contents
Renewable Resources
Design of Array CSRRs Band-Stop Filter . . . . . . . . . . . . . . . . . . . . . . . 87
Kada Becharef, Keltoum Nouri, Boubakar Seddik Bouazza,
Mahdi Damou, and Tayeb Habib Chawki Bouazza
Improvement of DTC with 24 Sectors of Induction Motor by Using
a Three-Level Inverter and Intelligent Hysteresis Controllers . . . . . . . . 99
Habib Benbouhenni, Rachid Taleb, and Fayçal Chabni
Energy Control Strategy Analysis of Hybrid Power Generation
System for Rural Saharan Community in Algeria . . . . . . . . . . . . . . . . . 108
Fadhila Fodhil, Abderrahmane Hamidat, and Omar Nadjemi
Improving the Electrical Stability by Wind Turbine and UPFC . . . . . . 121
Djamel Eddine Tourqui, Meryem Benakcha, and Tayeb Allaoui
Photo-Thermal Study of the Optical Properties of Multi-layer
Coatings Based on Black Pigmented Coatings/SiO2
Applied for Solar Absorber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Faouzi Haddad, Mustapha Hatti, R. Zaamoum, and Khadidja Rahmoun
Sahara Potential and Sustainable Development of Algeria:
A Thermal Experimental Study of Parabolic Trough Solar
Concentrator with New Perspectives of Solar Energy . . . . . . . . . . . . . . 143
Lahlour Rafik, Bellel Nadir, and Bouguetaia Nadia
Ensemble of Support Vector Methods to Estimate Global Solar
Radiation in Algeria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Nahed Zemouri and Hassen Bouzgou
Fuzzy-Direct Power Control of a Grid Connected Photovoltaic
System Associate with Shunt Active Power Filter . . . . . . . . . . . . . . . . . . 164
Sabir Ouchen, Achour Betka, Jean Paul Gaubert, Sabrina Abdeddaim,
and Farida Mazouz
Fuzzy Control of a Wind System Based on the DFIG . . . . . . . . . . . . . . 173
Farida Mazouz, Sebti Belkacem, Sabir Ouchen, Youcef Harbouche,
and Rachid Abdessemed
1
LAE, Laboratoire Architecture et Environnement,
Ecole Polytechnique d’Architecture et d’Urbanisme EPAU, Algiers, Algeria
otizouiar@yahoo.com
2
Ecole Polytechnique d’Architecture et d’Urbanisme, Algiers, Algeria
1 Introduction
The building sector is the largest energy consumer among the various economic sectors
in the world; its consumption represents 40% of the world’s energy, and therefore a
carbon dioxide emitter. Faced to this situation, and in order to respond to environmental
occupations of today, building must adapt. Taking into account regulatory requirements,
technical developments (the advent of technology and renewable energy sources), and
the new societal aspects (user behavior). This will give rise to an ecological and smart
management building. The smart building is today an element of major response to the
different modes of energy consumption as well as in terms of comfort and simplification
of inhabitant life. This can done only through development of new information and
communication technologies (ICT); and therefore the use of home automation equip‐
ment; from heating to ventilation, lighting, and shutter and blind management, which is
supposed to bring the best solutions to the performance of building.
First, our interest has focused on the subject of reasoned introduction of domotic and
multimedia technologies in architectural work, in order to have a house that simplifies
life to its inhabitants, a house where everything is comfortable, communicating, scalable,
autonomous, safe and economical. Our goal is to draw on the domain of home automa‐
tion for energy efficiency, which is a very important aspect in a smart residential
building. This is through integration of innovative constructive solutions based on the
smart management of produced and consumed energy. The research also reveals the
emergence of this concept in Algeria and it deepens about various difficulties and defi‐
ciencies, which caused this flagrant lack of this concept in our society.
The second part of our research focuses on attractiveness and energy efficiency in
the scale of building, using technologies, which seems to be an inescapable approach
all over the world. Algeria has taken note of these innovations by creating a technological
park, namely the Cyber Park of Sidi Abdellah in Algiers. In this sense, the present
research consists in understanding the different concepts and notions related to smart
buildings, the study of the interaction between occupant and his physical environment
with satisfaction level offered by this specific building. As well, as evaluate the experi‐
ence success of smart building as an actor of technological attractiveness and energy
saving.
2 Problematic
The main problems that arise for residential or tertiary smart buildings are the following:
“What are the architectural features and techniques that allow good energy
management in a smart residential building?” This is part of a current debate (fight
against global warming) and emanates from a basic fact problematic related to the
Algerian context, trying to verify whether smart residential buildings exist in Algeria,
and understand how the masters of works and promoters do for their designs and how
satisfied occupants with these smart buildings?
“Is the development of attractive smart building subject to constraints social
acceptability of technologies?” This concerns the lived experience of technological
buildings, specifically for the administrative seat block and multi-building -locator of
the Cyber Park of Sidi Abdellah, trying to verify the impact of automation degree on
the appropriation of their spaces by users.
3 Cases Study
4 Methodological Approach
The research begins with a state of knowledge on the different principles of smart build‐
ings and their implementation mechanisms and approaches throughout the world and
over time [3, 4, 6, 8, 9].
In the case of residential buildings, this theoretical step has been combined with
another phase, which is based on a fieldwork carried out in a qualitative dimension
focusing masters of works or companies realizing this type of construction in Algeria.
This, in order to situate themselves in relation to this type of architecture and make a
comparison with international references. The work is accompanied by two interviews
for the residents of this type of building as well as for the project managers; In order to
allow a better understanding of the subject.
Also in the case of the Cyber Park, the state of knowledge on so-called “smart or
technological” buildings and their attractive role in terms of comfort, telecommunica‐
tions, security and management has been combined with another phase, the analysis of
places in order to list the technical management system in the chosen buildings, and the
comfort facilities. Then a survey is carried out on a representative sample in order to
qualify the quality of use. Use of the post-occupational qualitative method (POE) is
adopted for this case as a tool for assessing performance of smart buildings and assessing
user satisfaction based essentially on impressions and perception of users on their work‐
space [5].
The survey in this part of the study takes place at the first two levels “indicator” and
“investigator”, without the “diagnostic” level. The pre-survey concerns the collection
of data on the state of the premises and this through guided tours. A documentary and
iconographic collection was compiled for each building. The second step is “mainte‐
nance”. It was from this database that the analysis was carried out. We then came up
with a grid of analysis by establishing the thematic categories on which the survey is
constructed (Fig. 1).
6 T. Ouahiba et al.
Fig. 1. Technology method used for qualitative analysis and description of the representative
sample
Smart Individual Housing in “Paradou, Algiers”: The site constraints have condi‐
tioned some decisions by designers, which has influenced the passive energy efficiency
of the house (implantation and orientation [10], Shape, openings, materials); After which
a few criteria of home automation are applied, either to the service of energy or comfort.
The Fig. 2 shows the percentage of the criteria used for home automation in the field of
energy efficiency and home automation for comfort. These criteria are deduced from the
theoretical part of our research (management of sunshades/blinds, control and manage‐
ment of heating, dual flow ventilation plant, access control, and technical detection,
sound.
Fig. 2. Passif and technological criteria applied in Smart individual housing in “Paradou, Algiers”
Smart Housing in Oran: The influence of northern orientation of rooms (less of natural
light) [10] has increased relatively the use of home automation for energy efficiency
(Fig. 3).
Smart Buildings and Occupants Satisfaction 7
Fig. 3. Passif and technological criteria applied in Smart individual housing in “Oran, Algeria”
Fig. 4. The criteria applied in smart residential building “Residence des pins” of the Lebanese
promotion in Chéraga
to know the various deficiencies and difficulties caused by this flagrant lack of this type
of concept in Algeria, among them there is the high cost of the home automation instal‐
lations and these techniques and also the housing crisis of which our country suffers
nowadays. However, for the time being there is no regulation in this sense, for a sustain‐
able and smart construction in its own sense.
First, for the Administration Building: The category that expressed its satisfaction
consists of engineers who are actors in the management of these buildings who are
optimistic for improvement of this system. Some occupants do not appreciate this type
of building in reference to their offices, it feels “bad”, this state is mainly related to
physical and psychological needs. This state of dissatisfaction mainly concerns the
feeling of confinement and the discomfort expressed by the occupants. On the other hand
most of the occupants are more or less satisfied with the comfort devices put in place
because according to the latter they ensure a better deliverability and a speed in the
exchanges. At this stage of analysis we can observe that dissatisfaction is linked to
factors concerning the envelope and the choice of the technical management system
(Fig. 5).
Result of the Level of Satisfaction of the Cooling and Heating Factor: The majority of
respondents were satisfied (54%), the rest were unsatisfactory (20%), moderately satis‐
fied (13%) and acceptable (13%). They consider that offices are not suitable or moder‐
ately suitable, because of the centralized system used. First, that is generalized for all
offices, it is the cold witch be regulated for all offices heating also, the occupant can not
intervene to satisfy his own need because the regularization for his office will be the
same for the neighboring office whereas the orientations of the offices are different. The
period of sunshine of offices in the day is different too. Adding to constant interruption
of electricity cause failures in the system whose absence of alternatives planned for this
kind of problem [1, 2].
The cooling control is very frequent, 53% of the respondents often intervene in the
regulation of the air-conditioning, while 47% are more satisfied with the centralized
control, but the control would always do it in their offices.
Same for heating 40% responded with thermal comfort, while 27% said it was
acceptable, 20% unsatisfactory and 13% moderately satisfactory. This dissatisfaction is
always argued by the constraints of centralized system control. The heating control
frequency is high with 87%, 47% “always” and 40% “often” (Fig. 6).
Fig. 6. Satisfaction level of the cooling and heating factor for the administration building
Level of Satisfaction of the Lighting Factor [1, 3]: Overall, the lighting was assessed
as “comfortable” and we noticed respondents’ appreciation of introduced intelligence
(presence detection). Indeed 60% often regulates the intensity, 27% never and 13%
rarely. Whereas other (40%) prefer to turn off the light, (47%) rarely and (13%) justify
what is satisfactory. These choices are justified by the good sunshine of the offices and
their orientations.
Level of Satisfaction of the Air Quality Factor: Air conditioning was designed as the
only source in offices, two-thirds of the occupants did not perceive it (73%) because of
10 T. Ouahiba et al.
unsatisfactory new air intake and low and non-automatic extraction, respondents pointed
out that the extraction of air requires a great power especially maintenance of the instal‐
lations what absent in these buildings. For this reason, the occupants seem to perceive
another source of ordinary aeration by the opening of the doors while the windows are
conceived to be closed.
Satisfaction Level of the Safety Factor: Almost all occupants expressed satisfaction at
being safe. However, 13% consider this supervision as a limitation of the freedoms of
the workers who have as objectives in these buildings to give their best for a better
deliverability. All occupants are sensitized to the subject of fire and they prefer to follow
the instructions of the services concerned by the intervention in case of danger, instead
of acting individually. We found that 40% of the respondents tried the manual fire system
arguing that this is the best way to test the operation of these. But sometimes this is
meant to stimulate a fire and make evacuation tests.
Fig. 7. Occupants trend level facing the energy consumption for the administration building
Level of Satisfaction: The category that says the type of building has no impact on work
performance is totally influenced by the physical urban setting of their workplace. The
urban context of these buildings is unpleasant, an unsatisfactory of the service and
accompanying equipment, and the difficulty of access to the site.
The overall satisfaction concerns the expectations linked mainly to the well being
lived in a framework of professional life offered by this type of architecture. The
Smart Buildings and Occupants Satisfaction 11
expressed dissatisfaction with power cuts seems to be accentuated among the people
who occupy the offices every hour of work, this disturbance results in delays and incon‐
veniences in performing the occupant’s spots.
The open space organization seems unfavorable in some because it limits privatiza‐
tion (Fig. 8).
Level of Satisfaction of the Heating and Cooling Factor: Heating and air conditioning
are acceptable. Complaints from occupants reveal the influence of the factors of office
volume, number of occupants and control over the satisfaction of the management
system and desired comfort (Fig. 9).
Fig. 9. Satisfaction level of cooling and heating factor for the multi-locator building
Lighting Factor Satisfaction Level [1, 3]: All the occupants are satisfied with the lighting
of the offices either natural or artificial. The need for lighting is high expressed, on
average 60% of respondents who turn off the light, 20% “never” while 20% expressed
their willingness to turn off the light. The active control question showed that 47%
(rarely) regulates intensity while 53% are satisfied with the intensity already
programmed.
12 T. Ouahiba et al.
Level of Satisfaction of the Air Quality Factor: The air quality in the surveyed office is
rated as acceptable (53%), (27%) feel comfortable, (13%) unsatisfactory and (7%)
moderately satisfactory.
Air conditioning is not the only source of ventilation in the offices, unlike offices of
the first building. As the air extraction is low so the return to the windows opening is
maintained to renew indoor air. Indeed, there was satisfaction with the occupants (80%)
nevertheless (20%) of people open the windows for no specific reason while the heating
or air conditioning is working (Fig. 10).
Fig. 10. Satisfaction level of lighting factor and air quality for multi tenant building
Safety Factor Satisfaction: All respondents are satisfied with the security system, they
pointed out that the safety factor is one of the criteria for which tenants chose to occupy
these offices. About fire, (20%) of the respondents are satisfied with the safety devices
against fire while (80%) of which (20%) prefers to go towards the rescue and 60% did
not answer the question.
Level of Satisfaction of the Internet Network: Overall satisfaction with the speed of the
Internet network introduced in the offices facilitated exchanges and communication and
minimized travel. However, electricity cuts are a major problem in the normal course
of work.
6 General Conclusion
Despite the difficulty of accessing examples of smart houses that rarely appear on
Algiers. Given the difficulty of having a large percentage of the sample, we were able
to conclude and understand that:
Smart Buildings and Occupants Satisfaction 13
References
3. Chenailler, H.: L’efficacité d’usage énergétique : pour une meilleure gestion de l’énergie
électrique intégrant les occupants dans les bâtiment, 280 p. Ph.D. thesis, electronic
department, National Polytechnic Institute of Grenoble (INPG), Grenoble (2012)
4. Dutreix, A.: Bioclimatisme et performances énergétiques des bâtiments, 239 p. Edition
Eyrolles, Bd Saint Germain, Paris (2010)
5. Fischer, G.N., Vischer, J.C.: Evaluation des environnements de travail: la méthode
diagnostique», 264 p. Montréal: Presse de l’Université de Montréal, and Brussels, DeBoek
(1998)
6. Gonzalo, R., Habermann, K.: Architecture et efficacité énergétique: principes de conception
et de construction, 215 p. BIRKHAUSER Edition, Boston, Berlin (2006)
7. Hamel, T.: Investigation sur les bâtiments résidentiels smarts en Algérie: L’étude de
l’efficacité énergétique de ces bâtiments, 72 p. Master memory in architecture at the
Polytechnic School of Architecture and Urbanism of Algiers, directed by Tizouiar, O,
November 2016
8. Jeuland, F-X.: La maison communicante. 306 p. 2nd edn. Eyrolles, Saint Germain, Paris
(2004)
9. Jeuland, F-X.: Réussir son installation domotique et multimédia. 381 p. 2nd edn. Eyrolles,
Saint Germain, Paris (2008)
10. Tizouiar, O.: Disponibilité de l’éclairage naturel en milieu urbain dense - investigation sur
les performances de puits de lumière naturelle, 185 p. Memory of magister in sciences in
architecture at the Polytechnic School of Architecture and Urbanism of Algiers, directed by
Pr Bensalem Rafik (2012)
Human Activity Recognition in Smart Home
Using Prior Knowledge Based
KNN-WSVM Model
1 Introduction
The growing population of elders in our society calls for a new approach in care giving
[1] to ensure the comfort of old people and because the healthcare infrastructures are
unlikely to accommodate the drastic growth of elderly population. Smart systems are
equipped with sensor networks able to automatically recognize activities about the
occupants and assist humans. They must be able to recognize the ongoing activities of
the users in order to suggest or take actions in an intelligent manner. Activity recog-
nition can be used to automatically monitor the activities of daily living (ADLs) of old
people such as cooking, brushing, dressing, cleaning, bathing and so on substantially
increased as well.
Feature extraction and classification are two key steps for activity recognition in a
smart home environment [2]. In this paper, we proposed a new hybrid classification
method KNN-WSVM using the K-Nearest Neighbors (KNN) [3] and Weighted Sup-
port Vector Machines (WSVM) [4]. The last is used to deal the class imbalance in the
training data due to the fact that people do not spend the same amount of time on the
different activities. We also integrated the prior knowledge [5] in this method to
improving the classification performances of KNN-WSVM. We introduced a new
feature: the temporal feature (TF) of activity when the activity is performed. Then we
added them to the existing features obtained by the collected wireless sensors data. The
‘Prepare breakfast’ and ‘Prepare dinner’ activities share the same model as they involve
the same set of object interactions. These two activities are distinguished by time of
taking place, i.e. ‘Prepare breakfast’ takes place in the morning hours and “Prepare
dinner” takes place in the afternoon or evening hours of the day.
Our paper addresses these issues and contributes on the following topics. Firstly,
we have presented the related works of human activity classification methods. Then, we
explained the proposed activity recognition method using the feature insertion using the
temporal attribute to recognize activities of daily living from binary sensor data. The
next section presents the experimental setup and discusses the results acquired
throughout a series of benchmark dataset [6, 7] constituted of highly imbalanced
datasets under different metrics. Finally, conclusions and future work are drawn in the
last section.
2 Related Works
The machine learning of the observed sensor patterns is usually done in a supervised
manner and requires a large annotated data sets recorded in different settings [6, 7]. For
this purpose, annotation of data for classification task can be performed in many
different ways, e.g. use of cameras [8], self-reporting approaches [7] and monitoring the
diary activity [6]. Several classification algorithms have been employed for ADL
recognition tasks [6, 8, 9], e.g. Hidden Markov model (HMM) [6], Conditional Ran-
dom Fields (CRF) [6], Linear Discriminant Analysis (LDA) [9], Bayes approach [8],
Support Vector Machine (SVM) [9] and its Soft-margin multiclass SVM extension
[10]. In [11], we developed a new classification method named PCA-LDA-WSVM
based on a combination of Principal Component Analysis (PCA), Linear Discriminant
Analysis (LDA) and Weighted Support Vector Machines (WSVM). We demonstrated
the ability of this method to achieve good improvement over the standard used methods
as HMM and CRF. This method resolved two problems result in a degradation of the
performance of activity recognition; the non-informative sequence features and the
class imbalance problem. In the next section, we present the proposed approach using
KNN-WSVM.
3 Proposed Approach
generated by KNN classifier. The second is the temporal feature (TF). This new dis-
tribution data is therefore used for learning and testing the multi-class weighted SVM
classifier using a new automated criterion for weighting the data. The choice of WSVM
like classifier [12, 13] in this context is motivated by previous work as shown in [11].
A generic diagram is shown in Fig. 1. Particularly, the initial dataset is divided into
training and test datasets.
Fig. 2. Temporal feature (TF) for TK26 M dataset. In red, the hour of beginning activity.
vectors in an n-dimensional feature space, each with a class label. In the KNN method,
the result of a new query is classified based on the majority of the KNN category. This
rule is usually called ‘voting KNN rule’. The classifiers do not use any model for fitting
and are only based on memory to store the feature vectors and class labels of the
training instances. They work based on the minimum distance from an unlabelled
vector (a test point) to the training instances to determine the K-nearest neighbors.
K (positive integer) is a user-defined constant. Usually Euclidean distance is used as the
distance metric.
The primal formulation of the margin in SVM maximizes margin 2/K(w, w) between
two classes and minimizes the amount of total misclassifications (training errors) ni as
follows:
X
m
min 1=2: Kðw; wÞ þ C ni
w;b;n
i¼1 ð1Þ
subject to yi ðwT /ðxi Þ þ bÞ 1 ni ; ni 0; i ¼ 1; . . . ; m
where w is normal to the hyperplane, b is the translation factor of the hyperplane to the
origin and /ð:Þ is a non-linear function which maps the input space into a feature space
defined by Kðxi ; xj Þ ¼ /ðxi ÞT /ðxj Þ.
We used the Gaussian kernel as follows: Kðx; yÞ ¼ exp kx yk2 =2r2 where r
is the width parameter. The construction of such functions is described by the Mercer
conditions [13]. The regularization parameter C is used to control the trade-off between
maximization of the margin width and minimizing the number of training error.
Solving the formulation dual of SVM [13] gives a decision function in the original
space for classifying a test point x 2 Rn
!
X
msv
f ðxÞ ¼ sgn ai yi Kðx; xi Þ þ b ð2Þ
i¼1