IT training computational intelligence in data mining (vol 3) jain, behera, mandal mohapatra 2014 12 11

Saibaba Reddy Vice Chancellor, VSSUT, Burla, Odisha, India Honorary General Chair Prof.. Rajib Mall, Ph.D., Professor and Head Department of Computer Science and Engineering, IIT Kharagp

Lakhmi C Jain

Himansu Sekhar Behera

Jyotsna Kumar Mandal

Durga Prasad Mohapatra

Smart Innovation, Systems and Technologies

Volume 33

Series editors

Robert J Howlett, KES International, Shoreham-by-Sea, UK

e-mail: rjhowlett@kesinternational.org

Lakhmi C Jain, University of Canberra, Canberra, Australia, and

University of South Australia, Adelaide, Australia

e-mail: Lakhmi.jain@unisa.edu.au

About this Series

The Smart Innovation, Systems and Technologies book series encompasses thetopics of knowledge, intelligence, innovation and sustainability The aim of theseries is to make available a platform for the publication of books on all aspects ofsingle and multi-disciplinary research on these themes in order to make the latestresults available in a readily-accessible form Volumes on interdisciplinaryresearch combining two or more of these areas is particularly sought

The series covers systems and paradigms that employ knowledge andintelligence in a broad sense Its scope is systems having embedded knowledgeand intelligence, which may be applied to the solution of world problems inindustry, the environment and the community It also focusses on the knowledge-transfer methodologies and innovation strategies employed to make this happeneffectively The combination of intelligent systems tools and a broad range ofapplications introduces a need for a synergy of disciplines from science,technology, business and the humanities The series will include conferenceproceedings, edited collections, monographs, handbooks, reference books, andother relevant types of book in areas of science and technology where smartsystems and technologies can offer innovative solutions

High quality content is an essential feature for all book proposals accepted forthe series It is expected that editors of all accepted volumes will ensure thatcontributions are subjected to an appropriate level of reviewing process and adhere

to KES quality principles

More information about this series at http://www.springer.com/series/8767

Lakhmi C Jain • Himansu Sekhar Behera

Jyotsna Kumar Mandal •

Durga Prasad Mohapatra

Himansu Sekhar Behera

Department of Computer Science

Kalyani UniversityNadia, West BengalIndia

Durga Prasad MohapatraDepartment of Computer Scienceand Engineering

National Institute of Technology RourkelaRourkela

India

Smart Innovation, Systems and Technologies

DOI 10.1007/978-81-322-2202-6

Library of Congress Control Number: 2014956493

Springer New Delhi Heidelberg New York Dordrecht London

© Springer India 2015

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.

Printed on acid-free paper

Springer (India) Pvt Ltd is part of Springer Science+Business Media (www.springer.com)

The First International Conference on“Computational Intelligence in Data Mining(ICCIDM-2014)” was hosted and organized jointly by the Department of ComputerScience and Engineering, Information Technology and MCA, Veer Surendra SaiUniversity of Technology, Burla, Sambalpur, Odisha, India between 20 and 21December 2014 ICCIDM is an international interdisciplinary conference coveringresearch and developments in the fields of Data Mining, Computational Intelli-gence, Soft Computing, Machine Learning, Fuzzy Logic, and a lot more More than

550 prospective authors had submitted their research papers to the conference.ICCIDM selected 192 papers after a double blind peer review process by experi-enced subject expertise reviewers chosen from the country and abroad The pro-ceedings of ICCIDM is a nice collection of interdisciplinary papers concerned invarious prolific research areas of Data Mining and Computational Intelligence Ithas been an honor for us to have the chance to edit the proceedings We haveenjoyed considerably working in cooperation with the International Advisory,Program, and Technical Committees to call for papers, review papers, andfinalizepapers to be included in the proceedings

This International Conference ICCIDM aims at encompassing a new breed ofengineers, technologists making it a crest of global success It will also educate theyouth to move ahead for inventing something that will lead to great success Thisyear’s program includes an exciting collection of contributions resulting from asuccessful call for papers The selected papers have been divided into thematicareas including both review and research papers which highlight the current focus

of Computational Intelligence Techniques in Data Mining The conference aims atcreating a forum for further discussion for an integrated information field incor-porating a series of technical issues in the frontier analysis and design aspects ofdifferent alliances in the relatedfield of Intelligent computing and others Thereforethe call for paper was on three major themes like Methods, Algorithms, and Models

in Data mining and Machine learning, Advance Computing and Applications.Further, papers discussing the issues and applications related to the theme of theconference were also welcomed at ICCIDM

v

The proceedings of ICCIDM have been released to mark this great day inICCIDM which is a collection of ideas and perspectives on different issues andsome new thoughts on various fields of Intelligent Computing We hope theauthor’s own research and opinions add value to it First and foremost are theauthors of papers, columns, and editorials whose works have made the conference agreat success We had a great time putting together this proceedings The ICCIDMconference and proceedings are a credit to a large group of people and everyoneshould be there for the outcome We extend our deep sense of gratitude to all fortheir warm encouragement, inspiration, and continuous support for making itpossible.

Hope all of us will appreciate the good contributions made and justify ourefforts

The theme and relevance of ICCIDM has attracted more than 550 researchers/academicians around the globe, which enabled us to select good quality papers andserve to demonstrate the popularity of the ICCIDM conference for sharing ideasand researchfindings with truly national and international communities Thanks toall who have contributed in producing such a comprehensive conferenceproceedings of ICCIDM

The organizing committee believes and trusts that we have been true to the spirit

of collegiality that members of ICCIDM value, even as maintaining an elevatedstandard as we have reviewed papers, provided feedback, and present a strong body

of published work in this collection of proceedings Thanks to all the members

of the Organizing committee for their heartfelt support and cooperation

It has been an honor for us to edit the proceedings We have enjoyed erably working in cooperation with the International Advisory, Program, andTechnical Committees to call for papers, review papers, andfinalize papers to beincluded in the proceedings

consid-We express our sincere thanks and obligations to the benign reviewers forsparing their valuable time and effort in reviewing the papers along with sugges-tions and appreciation in improvising the presentation, quality, and content of thisproceedings Without this commitment it would not be possible to have theimportant reviewer status assigned to papers in the proceedings The eminence

of these papers is an accolade to the authors and also to the reviewers who haveguided for indispensable perfection

We would like to gratefully acknowledge the enthusiastic guidance and tinuous support of Prof (Dr.) Lakhmi Jain, as and when it was needed as well asadjudicating on those difficult decisions in the preparation of the proceedings andimpetus to our efforts to publish this proceeding

con-Last but not the least, the editorial members of Springer Publishing deserve aspecial mention and our sincere thanks to them not only for making our dreamcome true in the shape of this proceedings, but also for its brilliant get-up andin-time publication in Smart, Innovation, System and Technologies, Springer

vii

I feel honored to express my deep sense of gratitude to all members of InternationalAdvisory Committee, Technical Committee, Program Committee, OrganizingCommittee, and Editorial Committee members of ICCIDM for their unconditionalsupport and cooperation.

The ICCIDM conference and proceedings are a credit to a large group of peopleand everyone should be proud of the outcome

Himansu Sekhar Behera

About the Conference

The International Conference on “Computational Intelligence in Data Mining”(ICCIDM-2014) has been established itself as one of the leading and prestigiousconference which will facilitate cross-cooperation across the diverse regionalresearch communities within India as well as with other International regionalresearch programs and partners Such an active dialogue and discussion amongInternational and National research communities is required to address many newtrends and challenges and applications of Computational Intelligence in thefield ofScience, Engineering and Technology ICCIDM 2014 is endowed with an oppor-tune forum and a vibrant platform for researchers, academicians, scientists, andpractitioners to share their original research findings and practical developmentexperiences on the new challenges and budding confronting issues

The conference aims to:

• Provide an insight into current strength and weaknesses of current applications aswell as researchfindings of both Computational Intelligence and Data Mining

• Improve the exchange of ideas and coherence between the various ComputationalIntelligence Methods

• Enhance the relevance and exploitation of data mining application areas for user as well as novice user application

end-• Bridge research with practice that will lead to a fruitful platform for the opment of Computational Intelligence in Data mining for researchers andpractitioners

devel-• Promote novel high quality research findings and innovative solutions to thechallenging problems in Intelligent Computing

• Make a tangible contribution to some innovative findings in the field of datamining

• Provide research recommendations for future assessment reports

ix

So, we hope the participants will gain new perspectives and views on currentresearch topics from leading scientists, researchers, and academicians around theworld, contribute their own ideas on important research topics like Data Mining andComputational Intelligence, as well as network and collaborate with their interna-tional counterparts.

Conference Committee

Patron

Prof E Saibaba Reddy

Vice Chancellor, VSSUT, Burla, Odisha, India

Honorary General Chair

Prof P K Dash, Ph.D., D.Sc., FNAE, SMIEEE, Director

Multi Disciplinary Research Center, S‘O’A University, India

Prof Lakhmi C Jain, Ph.D., M.E., B.E.(Hons), Fellow (Engineers Australia),University of Canberra, Canberra, Australia and University of South Australia,Adelaide, SA, Australia

Honorary Advisory Chair

Prof Shankar K Pal, Distinguished Professor

Indian Statistical Institute, Kolkata, India

General Chair

Prof Rajib Mall, Ph.D., Professor and Head

Department of Computer Science and Engineering, IIT Kharagpur, India

xi

Program Chairs

Dr Sukumar Mishra, Ph.D., Professor

Department of EE, IIT Delhi, India

Prof H.S Behera, Reader, Department of Computer Science Engineering andInformation Technology, Veer Surendra Sai University of Technology, Burla,Odisha, India

Prof J.K Mandal, Professor, Department of Computer Science and Engineering,University of Kalyani, Kolkata, India

Prof D.P Mohapatra, Associate Professor, Department of Computer Science andEngineering, NIT, Rourkela, Odisa, India

International Advisory Committee

Prof C.R Tripathy (VC, Sambalpur University)

Prof B.B Pati (VSSUT, Burla)

Prof A.N Nayak (VSSUT, Burla)

Prof S Yordanova (STU, Bulgaria)

Prof P Mohapatra (University of California)

Prof S Naik (University of Waterloo, Canada)

Prof S Bhattacharjee (NIT, Surat)

Prof G Saniel (NIT, Durgapur)

Prof K.K Bharadwaj (JNU, New Delhi)

Prof Richard Le (Latrob University, Australia)

Prof K.K Shukla (IIT, BHU)

Prof G.K Nayak (IIIT, BBSR)

Prof S Sakhya (TU, Nepal)

Prof A.P Mathur (SUTD, Singapore)

Prof P Sanyal (WBUT, Kolkata)

Prof Yew-Soon Ong (NTU, Singapore)

Prof S Mahesan (Japfna University, Srilanka)

Prof B Satapathy (SU, SBP)

Prof G Chakraborty (IPU, Japan)

Prof T.S Teck (NU, Singapore)

Prof P Mitra (P.S University, USA)

Prof A Konar (Jadavpur University)

Prof S Das (Galgotias University)

Prof A Ramanan (UJ, Srilanka)

Prof Sudipta Mohapatra, (IIT, KGP)

Prof P Bhattacharya (NIT, Agaratala)

Prof N Chaki (University of Calcutta)

Dr J.R Mohanty, (Registrar, VSSUT, Burla)

Prof M.N Favorskaya (SibSAU)

Mr D Minz (COF, VSSUT, Burla)

Prof L.M Patnaik (DIAT, Pune)

Prof G Panda (IIT, BHU)

Prof S.K Jena (NIT, RKL)

Prof V.E Balas (University of Arad)

Prof R Kotagiri (University of Melbourne)

Prof B.B Biswal (NIT, RKL)

Prof Amit Das (BESU, Kolkata)

Prof P.K Patra (CET, BBSR)

Prof N.G.P.C Mahalik (California)

Prof D.K Pratihar (IIT, KGP)

Prof A Ghosh (ISI, Kolkata)

Prof P Mitra (IIT, KGP)

Prof P.P Das (IIT, KGP)

Prof M.S Rao (JNTU, HYD)

Prof A Damodaram (JNTU, HYD)

Prof M Dash (NTU, Singapore)

Prof I.S Dhillon (University of Texas)

Prof S Biswas (IIT, Bombay)

Prof S Pattnayak (S‘O’A, BBSR)

Prof M Biswal (IIT, KGP)

Prof Tony Clark (M.S.U, UK)

Prof Sanjib ku Panda (NUS)

Prof G.C Nandy (IIIT, Allahabad)

Prof R.C Hansdah (IISC, Bangalore)

Prof S.K Basu (BHU, India)

Prof P.K Jana (ISM, Dhanbad)

Prof P.P Choudhury (ISI, Kolkata)

Prof H Pattnayak (KIIT, BBSR)

Prof P Srinivasa Rao (AU, Andhra)

International Technical Committee

Dr Istvan Erlich, Ph.D., Chair Professor, Head

Department of EE and IT, University of DUISBURG-ESSEN, Germany

Dr Torbjørn Skramstad, Professor,

Department of Computer and System Science, Norwegian University

of Science and Technology, Norway

Dr P.N Suganthan, Ph.D., Associate Professor

School of EEE, NTU, Singapore

Prof Ashok Pradhan, Ph.D., Professor

Department of EE, IIT Kharagpur, India

Head, Department of IT, B.I.T, Meshra, India

Dr Amit Saxena, Ph.D., Professor

Head, Department of CS and IT, CU, Bilashpur, India

Dr Sidhartha Panda, Ph.D., Professor

Department of EEE, VSSUT, Burla, Odisha, India

Dr Swagatam Das, Ph.D., Associate Professor

Indian Statistical Institute, Kolkata, India

Dr Chiranjeev Kumar, Ph.D., Associate Professor and Head

Department of CSE, Indian School of Mines (ISM), Dhanbad

Dr B.K Panigrahi, Ph.D., Associate Professor

Department of EE, IIT Delhi, India

Dr A.K Turuk, Ph.D., Associate Professor

Head, Department of CSE, NIT, RKL, India

Dr S Samantray, Ph.D., Associate Professor

Department of EE, IIT BBSR, Odisha, India

Dr B Biswal, Ph.D., Professor

Department of ETC, GMRIT, A.P., India

Dr Suresh C Satpathy, Professor, Head

Department of Computer Science and Engineering, ANITS, AP, India

Dr S Dehuri, Ph.D., Associate Professor

Department of System Engineering, Ajou University, South Korea

Dr B.B Mishra, Ph.D., Professor,

Department of IT, S.I.T, BBSR, India

Dr G Jena, Ph.D., Professor

Department of CSE, RIT, Berhampu, Odisha, India

Dr Aneesh Krishna, Assistant Professor

Department of Computing, Curtin University, Perth, Australia

Dr Ranjan Behera, Ph.D., Assistant Professor

Department of EE, IIT, Patna, India

Dr A.K Barisal, Ph.D., Reader

Department of EE, VSSUT, Burla, Odisha, India

Dr R Mohanty, Reader

Department of CSE, VSSUT, Burla

Conference Steering Committee

Publicity Chair

Prof A Rath, DRIEMS, Cuttack

Prof B Naik, VSSUT, Burla

Mr Sambit Bakshi, NIT, RKL

Logistic Chair

Prof S.P Sahoo, VSSUT, Burla

Prof S.K Nayak, VSSUT, Burla

Prof D.C Rao, VSSUT, Burla

Prof K.K Sahu, VSSUT, Burla

Organizing Committee

Prof D Mishra, VSSUT, Burla

Prof J Rana, VSSUT, Burla

Prof P.K Pradhan, VSSUT, Burla

Prof P.C Swain, VSSUT, Burla

Prof P.K Modi, VSSUT, Burla

Prof S.K Swain, VSSUT, Burla

Prof P.K Das, VSSUT, Burla

Prof P.R Dash, VSSUT, Burla

Prof P.K Kar, VSSUT, Burla

Prof U.R Jena, VSSUT, Burla

Prof S.S Das, VSSUT, Burla

Prof Sukalyan Dash, VSSUT, Burla

Prof D Mishra, VSSUT, Burla

Prof S Aggrawal, VSSUT, Burla

Prof R.K Sahu, VSSUT, Burla

Prof M Tripathy, VSSUT, Burla

Prof K Sethi, VSSUT, Burla

Prof B.B Mangaraj, VSSUT, Burla

Prof M.R Pradhan, VSSUT, Burla

Prof S.K Sarangi, VSSUT, Burla

Prof N Bhoi, VSSUT, Burla

Prof J.R Mohanty, VSSUT, Burla

Prof Sumanta Panda, VSSUT, Burla

Prof A.K Pattnaik, VSSUT, Burla

Prof S Panigrahi, VSSUT, Burla

Prof S Behera, VSSUT, Burla

Prof M.K Jena, VSSUT, Burla

Prof S Acharya, VSSUT, Burla

Prof S Kissan, VSSUT, Burla

Prof S Sathua, VSSUT, Burla

Prof E Oram, VSSUT, Burla

Dr M.K Patel, VSSUT, Burla

Mr N.K.S Behera, M.Tech Scholar

Mr T Das, M.Tech Scholar

Mr S.R Sahu, M.Tech Scholar

Mr M.K Sahoo, M.Tech Scholar

Prof J.V.R Murthy, JNTU, Kakinada

Prof G.M.V Prasad, B.V.CIT, AP

Prof S Pradhan, UU, BBSR

Prof P.M Khilar, NIT, RKL

Prof Murthy Sharma, BVC, AP

Prof M Patra, BU, Berhampur

Prof M Srivastava, GGU, Bilaspur

Prof P.K Behera, UU, BBSR

Prof B.D Sahu, NIT, RKL

Prof S Baboo, Sambalpur University

Prof Ajit K Nayak, S‘O’A, BBSR

Prof Debahuti Mishra, ITER, BBSR

Prof S Sethi, IGIT, Sarang

Prof C.S Panda, Sambalpur University

Prof N Kamila, CVRCE, BBSR

Prof H.K Tripathy, KIIT, BBSR

Prof S.K Sahana, BIT, Meshra

Prof Lambodar Jena, GEC, BBSR

Prof R.C Balabantaray, IIIT, BBSR

Prof D Gountia, CET, BBSR

Prof Mihir Singh, WBUT, Kolkata

Prof A Khaskalam, GGU, Bilaspur

Prof Sashikala Mishra, ITER, BBSR

Prof D.K Behera, TAT, BBSR

Prof Shruti Mishra, ITER, BBSR

Prof H Das, KIIT, BBSR

Mr Sarat C Nayak, Ph.D Scholar

Mr Pradipta K Das, Ph.D Scholar

Mr G.T Chandrasekhar, Ph.D Scholar

Mr P Mohanty, Ph.D Scholar

Mr Sibarama Panigrahi, Ph.D Scholar

Mr A.K Bhoi, Ph.D Scholar

Mr T.K Samal, Ph.D Scholar

Mr Ch Ashutosh Swain, MCA

Mr Nrusingh P Achraya, MCA

Mr Devi P Kanungo, M.Tech Scholar

Mr M.K Sahu, M.Tech Scholar

Conceptual Modeling of Non-Functional Requirements

from Natural Language Text 1

S Abirami, G Shankari, S Akshaya and M Sithika

A Genetic Algorithm Approach for Multi-criteria Project

Selection for Analogy-Based Software Cost Estimation 13Sweta Kumari and Shashank Pushkar

A Service-Oriented Architecture (SOA) Framework Component

for Verification of Choreography 25Prachet Bhuyan, Abhishek Ray and Durga Prasad Mohapatra

Migration of Agents in Artificial Agent Societies: Framework

and State-of-the-Art 37Harjot Kaur, Karanjeet Singh Kahlon and Rajinder Singh Virk

Automated Colour Segmentation of Malaria Parasite

with Fuzzy and Fractal Methods 53M.L Chayadevi and G.T Raju

Syllable Based Concatenative Synthesis for Text

to Speech Conversion 65

S Ananthi and P Dhanalakshmi

Kannada Stemmer and Its Effect on Kannada Documents

Classification 75

N Deepamala and P Ramakanth Kumar

Boosting Formal Concept Analysis Based Definition Extraction

via Named Entity Recognition 87G.S Mahalakshmi and A.L Agasta Adline

xix

Harmonic Reduction Using PV Fuzzy with Inverter 99Subha Darsini Misra, Tapas Mohapatra and Asish K Nanda

Experimental Validation of a Fuzzy Model for Damage

Prediction of Composite Beam Structure 109Deepak K Agarwalla, Amiya K Dash and Biswadeep Tripathy

A Comparative View of AoA Estimation in WSN Positioning 123Sharmilla Mohapatra, Sasmita Behera and C.R Tripathy

Multi-strategy Based Matching Technique for Ontology

Integration 135

S Kumar and V Singh

Impact of Fuzzy Logic Based UPFC Controller on Voltage Stability

of a Stand-Alone Hybrid System 149Asit Mohanty and Meera Viswavandya

A Novel Modified Apriori Approach for Web

Document Clustering 159Rajendra Kumar Roul, Saransh Varshneya, Ashu Kalra

and Sanjay Kumar Sahay

Sliding-Window Based Method to Discover High Utility

Patterns from Data Streams 173Chiranjeevi Manike and Hari Om

Important Author Analysis in Research Professionals’

Relationship Network Based on Social Network Analysis Metrics 185Manoj Kumar Pandia and Anand Bihari

Predicting Software Development Effort Using Tuned

Artificial Neural Network 195H.S Hota, Ragini Shukla and S Singhai

Computer Vision Based Classification of Indian Gujarat-17

Rice Using Geometrical Features and Cart 205Chetna V Maheshwari, Niky K Jain and Samrat Khanna

A Novel Feature Extraction and Classification Technique

for Machine Learning Using Time Series and Statistical

Approach 217R.C Barik and B Naik

Implementation of 32-Point FFT Processor for OFDM System 229

G Soundarya, V Jagan Naveen and D Tirumala Rao

Performance Analysis of Microstrip Antenna Using GNU

Radio with USRP N210 239

S Santhosh, S.J Arunselvan, S.H Aazam, R Gandhiraj

and K.P Soman

Selection of Control Parameters of Differential Evolution

Algorithm for Economic Load Dispatch Problem 251Narendra Kumar Yegireddy, Sidhartha Panda, Umesh kumar Rout

and Rama Kishore Bonthu

Wireless QoS Routing Protocol (WQRP): Ensuring Quality

of Service in Mobile Adhoc Networks 261Nishtha Kesswani

CS-ATMA: A Hybrid Single Channel MAC Layer Protocol

for Wireless Sensor Networks 271Sonali Mishra, Rakesh Ranjan Swain, Tushar Kanta Samal

and Manas Ranjan Kabat

Energy Efficient Reliable Data Delivery in Wireless Sensor

Networks for Real Time Applications 281Prabhudutta Mohanty, Manas Ranjan Kabat

and Manoj Kumar Patel

A Rule Induction Model Empowered by Fuzzy-Rough Particle

Swarm Optimization Algorithm for Classification

of Microarray Dataset 291Sujata Dash

Bird Mating Optimization Based Multilayer Perceptron

for Diseases Classification 305N.K.S Behera, A.R Routray, Janmenjoy Nayak and H.S Behera

Comparison and Prediction of Monthly Average Solar Radiation

Data Using Soft Computing Approach for Eastern India 317Sthitapragyan Mohanty, Prashanta Kumar Patra and Sudhansu S Sahoo

Effect of Thermal Radiation on Unsteady Magnetohydrodynamic

Free Convective Flow in Vertical Channel in Porous Medium 327J.P Panda

A Fuzzy Based Approach for the Selection of Software

Testing Automation Framework 335Mohd Sadiq and Fahamida Firoze

Time Frequency Analysis and Classification of Power Quality

Events Using Bacteria Foraging Algorithm 345

S Jagadeesh and B Biswal

Non-Stationary Signal Analysis Using Time Frequency

Transform 353

M Kasi Subrahmanyam and Birendra Biswal

Multivariate Linear Regression Model for Host Based

Intrusion Detection 361Sunil Kumar Gautam and Hari Om

CBDF Based Cooperative Multi Robot Target Searching

and Tracking Using BA 373Sanjeev Sharma, Chiranjib Sur, Anupam Shukla and Ritu Tiwari

A Highly Robust Proxy Enabled Overload Monitoring

System (P-OMS) for E-Business Web Servers 385S.V Shetty, H Sarojadevi and B Sriram

Short Term Hydro Thermal Scheduling Using Invasive

Weed Optimization Technique 395A.K Barisal, R.C Prusty, S.S Dash and S.K Kisan

Contourlet Transform Based Feature Extraction Method

for Finger Knuckle Recognition System 407

K Usha and M Ezhilarasan

CDM Controller Design for Non-minimum Unstable Higher

Order System 417T.V Dixit, Nivedita Rajak, Surekha Bhusnur and Shashwati Ray

Model Based Test Case Prioritization Using Association

Rule Mining 429Arup Abhinna Acharya, Prateeva Mahali and Durga Prasad Mohapatra

Rank Based Ant Algorithm for 2D-HP Protein Folding 441

N Thilagavathi and T Amudha

Boolean Operations on Free Form Shapes in a Level

Set Framework 453V.R Bindu and K.N Ramachandran Nair

Feature Extraction and Recognition of Ancient Kannada

Epigraphs 469

A Soumya and G Hemantha Kumar

eKMP: A Proposed Enhancement of KMP Algorithm 479Nitashi Kalita, Chitra, Radhika Sharma and Samarjeet Borah

Acquisition and Analysis of Robotic Data Using Machine

Learning Techniques 489Shivendra Mishra, G Radhakrishnan, Deepa Gupta

and T.S.B Sudarshan

Maximum Likelihood DOA Estimation in Wireless Sensor

Networks Using Comprehensive Learning Particle Swarm

Optimization Algorithm 499Srinivash Roula, Harikrishna Gantayat, T Panigrahi and G Panda

Performance Evaluation of Some Clustering Indices 509Parthajit Roy and J.K Mandal

A New (n, n) Blockcipher Hash Function Using Feistel Network:

Apposite for RFID Security 519Atsuko Miyaji and Mazumder Rashed

Efficient Web Log Mining and Navigational Prediction

with EHPSO and Scaled Markov Model 529Kapil Kundra, Usvir Kaur and Dheerendra Singh

View Invariant Human Action Recognition Using

Improved Motion Descriptor 545

M Sivarathinabala, S Abirami and R Baskaran

Visual Secret Sharing of Color Image Using Extended

Asmuth Bloom Technique 555

L Jani Anbarasi, G.S Anadha Mala and D.R.L Prassana

Enhancing Teaching-Learning Professional Courses

via M-Learning 563

V Sangeetha Chamundeswari and G.S Mahalakshmi

Evaluation of Fitness Functions for Swarm Clustering

Applied to Gene Expression Data 571P.K Nizar Banu and S Andrews

Context Based Retrieval of Scientific Publications

via Reader Lens 583G.S Mahalakshmi, R Siva and S Sendhilkumar

Numerical Modeling of Cyclone-Generated Wave Around

Offshore Objects 597Subrata Bose and Gunamani Jena

Cooperative Resource Provisioning for Futuristic

Cloud Markets 607Geetika Mudali, Manas Ranjan Patra, K Hemant K Reddy

and Diptendu S Roy

Frequency Based Inverse Damage Assessment Technique

Using Novel Hybrid Neuro-particle Swarm Optimization 617Bharadwaj Nanda, R Anand and Dipak K Maiti

A Hybrid CRO-K-Means Algorithm for Data Clustering 627Sibarama Panigrahi, Balaram Rath and P Santosh Kumar

An Approach Towards Most Cancerous Gene Selection

from Microarray Data 641Sunanda Das and Asit Kumar Das

An Ordering Policy with Time-Proportional Deterioration,

Linear Demand and Permissible Delay in Payment 649Trailokyanath Singh and Hadibandhu Pattanayak

Computation of Compactly Supported Biorthogonal Riesz

Basis of Wavelets 659Mahendra Kumar Jena and Manas Ranjan Mishra

Data Mining Approach for Modeling Risk Assessment

in Computational Grid 673Sara Abdelwahab and Ajith Abraham

Degree of Approximation of Conjugate Series of a Fourier

Series by Hausdroff and Norlund Product Summability 685Sunita Sarangi, S.K Paikray, M Dash, M Misra and U.K Misra

A Novel Approach for Intellectual Image Retrieval Based

on Image Content Using ANN 693Anuja Khodaskar and Sidharth Ladhake

Hierarchical Agents Based Fault-Tolerant

and Congestion-Aware Routing for NoC 705Chinmaya Kumar Nayak, Satyabrata Das and Himansu Sekhar Behera

Author Index 715

Editors ’ Biography

Prof Lakhmi C Jainis with the Faculty of Education, Science, Technology andMathematics at the University of Canberra, Australia and University of SouthAustralia, Australia He is a Fellow of the Institution of Engineers, Australia.Professor Jain founded the Knowledge-Based Intelligent Engineering System(KES) International, a professional community for providing opportunities forpublication, knowledge exchange, cooperation, and teaming Involving around5,000 researchers drawn from universities and companies worldwide, KES facili-tates international cooperation and generates synergy in teaching and research KESregularly provides networking opportunities for the professional communitythrough one of the largest conferences of its kind in the area of KES His interestsfocus on artificial intelligence paradigms and their applications in complex systems,security, e-education, e-healthcare, unmanned air vehicles, and intelligent agents.Prof Himansu Sekhar Behera is working as a Reader in the Department ofComputer Science Engineering and Information Technology, Veer Surendra SaiUniversity of Technology (VSSUT) (A Unitary Technical University, Established

by Government of Odisha), Burla, Odisha He has received M.Tech in ComputerScience and Engineering from N.I.T, Rourkela (formerly R.E.C., Rourkela) andDoctor of Philosophy in Engineering (Ph.D.) from Biju Pattnaik University ofTechnology (BPUT), Rourkela, Government of Odisha respectively He has pub-lished more than 80 research papers in various international journals and confer-ences, edited 11 books and is acting as a member of the editorial/reviewer board ofvarious international journals He is proficient in the field of Computer ScienceEngineering and served in the capacity of program chair, tutorial chair, and acted asadvisory member of committees of many national and international conferences.His research interest includes Data Mining and Intelligent Computing He isassociated with various educational and research societies like OITS, ISTE, IE,ISTD, CSI, OMS, AIAER, SMIAENG, SMCSTA, etc He is currently guidingseven Ph.D scholars

xxvii

Prof Jyotsna Kumar Mandalis working as Professor in Computer Science andEngineering, University of Kalyani, India Ex-Dean Faculty of Engineering,Technology and Management (two consecutive terms since 2008) He has 26 years

of teaching and research experiences He was Life Member of Computer Society ofIndia since 1992 and life member of Cryptology Research Society of India, member

of AIRCC, associate member of IEEE and ACM His research interests includeNetwork Security, Steganography, Remote Sensing and GIS Application, ImageProcessing, Wireless and Sensor Networks Domain Expert of Uttar Banga KrishiViswavidyalaya, Bidhan Chandra Krishi Viswavidyalaya for planning and inte-gration of Public domain networks He has been associated with national andinternational journals and conferences The total number of publications to hiscredit is more than 320, including 110 publications in various international journals.Currently, he is working as Director, IQAC, Kalyani University

Prof Durga Prasad Mohapatra received his Ph.D from Indian Institute ofTechnology Kharagpur and is presently serving as an Associate Professor in NITRourkela, Odisha His research interests include software engineering, real-timesystems, discrete mathematics, and distributed computing He has published morethan 30 research papers in these fields in various international Journals and con-ferences He has received several project grants from DST and UGC, Government

of India He received the Young Scientist Award for the year 2006 from OrissaBigyan Academy He has also received the Prof K Arumugam National Awardand the Maharashtra State National Award for outstanding research work in Soft-ware Engineering for the years 2009 and 2010, respectively, from the IndianSociety for Technical Education (ISTE), New Delhi He is nominated to receive theBharat Shiksha Ratan Award for significant contribution in academics awarded bythe Global Society for Health and Educational Growth, Delhi

Conceptual Modeling of Non-Functional

Requirements from Natural Language

Text

S Abirami, G Shankari, S Akshaya and M Sithika

Abstract The conceptual model is an intermediate model that represents theconcepts (entities), attributes and their relationships that aids in the visualization ofrequirements In literature, explicit design elements such as concept, attribute,operations, relationships are extracted as functional requirements and they arerepresented in the conceptual model The conceptual model cannot be completeunless Non-Functional Requirements are included This problem has motivated us

to consider both functional and Non Functional Requirements in this research forthe development of a conceptual model Therefore, this paper presents a frameworkfor the development of conceptual model by designing a classifier which segregatesthe functional and Non Functional Requirements (NFR) from the requirementsautomatically Later, these requirements are transformed to the conceptual modelwith explicit visualization of NFR using design rules In addition, the results of thismodel are also validated against the standard models

Keywords Conceptual modeling
Non-functional requirements
Natural languageprocessing
NFR classifier

S Abirami ( &)
G Shankari
S Akshaya
M Sithika

Department of Information Science and Technology, College of Engineering,

Anna University, Chennai, India

L.C Jain et al (eds.), Computational Intelligence in Data Mining - Volume 3,

Smart Innovation, Systems and Technologies 33, DOI 10.1007/978-81-322-2202-6_1

1

1 Introduction

Automated Requirement Analysis is a kind of Information Extraction (IE) task inthe domain of Natural Language Processing (NLP) In fact, it is the most significantphase of Software Development Life Cycle (SDLC) [1] Errors caused during thisphase can be quite expensive tofix it later Requirements Gathering and Analysis isthe most critical phase of the SDLC and this could greatly benefit through theirautomation as said by VidhuBala et al [2] Automation of conceptual model notonly saves time but also helps in visualizing the problem Conceptual modelcontains only domain entities and attributes according to Larman [3]

All the researches in the literature contribute towards the extraction of functionrequirements from natural language text In addition to functional requirements,Non-Functional Requirements (NFR) should also be considered while creating aconceptual model This is because the non-satisfaction of NFRs has been consid-ered as one of the main reason for the failure of various software projects according

to Saleh [4] Therefore, we have proposed a system in this research which extractsNon-Functional Requirements with functional requirements automatically to makethe conceptual model a complete one The novelty of this paper lies in the devel-opment of the conceptual model named CMFNR thereby extracting and visualizingNon-Functional Requirements too

This paper is organized as follows: Sect.2discusses about the related works andSect.3describes our system of CM-NFR whereas Sect.4discusses the performanceobtained by our system and Sect.5 concludes the paper

2 Related Work

Here, literature survey has been done in two areas namely explicit functionalrequirements modeling and Non-Functional Requirements modeling Text mininghas been used to assist requirement analysis in [5] where in the responsibilities werefirst identified using POS tags and then grouped semantically using clusteringalgorithms But this approach requires human interaction to identify the conceptualclasses Later, Use case diagram, Conceptual model, Collaboration diagram andDesign class diagram were generated using a tool called UMGAR (UML ModelGenerator from Analysis of Requirements) in [6] But, it requires both requirementsspecification and use case specification in active voice form to generate the UMLdiagrams Requirements and use cases were processed to produce classes, attributesand functional modules using DAT (Design Assistant Tool) in [7]

Tools like NLARE have been developed in literature to evaluate the quality ofrequirements like atomicity, ambiguity and completeness in [8] Significant work

in automating the conceptual model creation has been done in [2] But theabove system assumes that the input is grammatically correct and handles onlyexplicit requirements Non-Functional Requirements (NFR) are not handled here

Taking Use Case Description (UCD) written in natural language as input, a parsedUCD and a UML class model were developed in [9], using a tool called Class-Gen.Parsed Use Case Description were subjected to rules to identify a list of candidateclasses and relations But the classes extracted should be refined by a human expertsince the rules were not complete All the above models discussed in the literatureconcentrate on functional requirements and they lack on the extraction of Non-Functional Requirements (NFR).

A novel approach to automatically transform NL specification of softwareconstraints to OCL constraints was done by Bajwa [10] According to UML, a note

is used to represent a NFR But this is not sufficient Non-Functional Requirementscomes in the second conceptual model of the class Name of the class corresponded

to the NFR being modeled and the entire requirement was specified in a note usingOCL (Object Constraint Language) Hence, a class diagram can be used asextension to model NFR [4] in an effective way As a result, in this research, weintend to develop an extended class diagram which could visualize the Non-Functional Requirements extracted automatically

3 Conceptual Model-Non functional Requirements

3.2 NFR Classi fication

NFR classifier uses a set of pre-defined knowledge base to classify the list ofsentences into two lists namely functional and Non-Functional Requirements Theknowledge base used by the classifier is specified below

3.2.1 Extracting Specific Non-functional Requirements and Its

Attributes

The classification of the NFR into a specific category namely performance, cost andinteroperability is achieved by the following rules that are stated at a Bird’s-eyeview level The novelty of our system lies in the construction of CM-NFR whichextracts Non-Functional Requirements automatically from requirements text visu-alizes both functional and Non-Functional Requirements in a complete way Syn-tactic rules used by the NFR Classifier to discriminate the Non-FunctionalRequirements from requirements has been listed below:

Classifier Rule 1: Whenever time is preceded by ‘at least within’, ‘atmost within’,

‘within’ clauses and the sentence contains the terms ‘throughput’ or ‘responsetime’, then it is a non-functional performance requirement

Example: The system should retrieve the results within 3 s

The system should have very less response time

Fig 1 Architecture of conceptual model generator

Classifier Rule 2: Whenever a sentence contains memory units and terms like

‘memory’, ‘storage’, ‘space’, then it is a non-functional memory requirement.Example: The application should not consume a memory space of more than 2 GB.Classifier Rule 3: Whenever a sentence contains the phrase ‘/second’ or ‘/minute’

or ‘per second’ etc then it is a non-functional performance requirement thatemphasizes speed

Example: the system should handle 100 transactions/s

Classifier Rule 4: Whenever a number is followed by rs, $, yen, pounds or anymonetary unit then it is a non-functional cost requirement

Classifier Rule 5: Whenever a sentence contains the stem ‘interoperate’ then it is anon-functional interoperability requirement

Example: The application shall properly interoperate with the legacy CustomerOracle database

The above rules were formulated after analyzing various requirements documentdone on number of projects in the development of software systems Sentenceswhich satisfy the classifier rules would be segregated either as functional or Non-Functional Requirements

3.3 Deep Parsing

The output of the previous module (functional and Non-Functional Requirements)

isfirst subjected to POS tagging using Stanford-POS tagger [12] to get POS tags(Part-Of-Speech) and then it is subjected to deep parsing using Stanford-parser toget typed dependencies and a parse tree

3.4 Syntactic Analysis

The POS tagged functional requirements and their typed dependencies are used bythe following rules to extract the required design elements as done in the conceptualmodeling [2] later The class rule, attribute and relation rules which were used inthis research to extract design elements specified explicitly has been identified andused over the POS tagged sentences (both functional and non functional) as listedbelow They are illustrated with an example for each in the following subsections:

3.4.1 Rules Used for Class Generation

Class Rule 1: Any Noun that appears as a subject is always a class [2]

Example: A bank owns an ATM

Class Rule 2: Nouns are always converted to their singular form This is done byobtaining the stemmed form of the word

Class Rule 3: Nouns occurring as objects participate in relations, but are notcreated as classes explicitly [2].

Example: The user presses the button

Class Rule 4: Gerunds are created as classes [9]

Gerund forms of verbs are treated as class names

Examples: Borrowing is processed by the staff Identified classes: Borrowing

3.4.2 Rules Used for Attribute Generation

Attribute Rule 1: A noun phrase which follows the phrase ‘identified by’, ognized by’ indicates presence of an attribute [9]

‘rec-Example: An employee is identified by employee id → Employee has attributeemployee id

Attribute Rule 2: An intransitive verb with an adverb may signify an attribute [9].Example: the train arrives in the morning at 8 AM.→ Train has an attribute, time.Attribute Rule 3: A classifiable adjective, either in the predicate form or inattributive form signifies an attribute [2]

Example: The red button glows when pressed.→ Button has an attribute, color.Attribute Rule 4: possessive apostrophe signifies an attribute [9]

Example: employee’s name is save → Name is an attribute of the employee.Attribute Rule 5: The genitive case when referring to a relationship of possessoruses the‘of’ construction [9]

Example: The name of the employee is printed → Name is an attribute of theemployee

3.4.3 Rules Used for Relation Generation

Relation rule 1: A transitive verb is a candidate for a relationship [9]

A transitive verb links a subject and an object If the object translates to a class, theverb becomes a relation

Example: The banker issues a cheque→ Relation (banker, 1, cheque, 1, issues).Relation rule 2: A verb with a preposition is a candidate for a relationship [2]

A verb that contains a prepositional object linked with a transitive verb along with apreposition, combines with the preposition to form a relationship

Example: The cheque is sent to the bank.→ sent _to is the relation

Relation rule 3: A sentence of the form‘the r of a is b’ [9]

Relation rule 4: A sentence of the form‘a is the r of b’ [2]

Example: The bank of the customer is SBI→ Relation (bank, 1, customer, 1, SBI).Example: SBI is the bank of the customer→ Relation (bank, 1, customer, 1, SBI)

In the above sentences, r is the relation that relates class a to class b The above tworules are examples of the relation being in the form of a noun, rather than a verb

3.4.4 Rules Used for Operation Generation

Operation rule 1: An intransitive verb is an operation [9]

An intransitive verb is a verb that does not need an object

Example: The laptop hibernates.→ hibernate is an operation of laptop

Operation rule 2: A verb that relates an entity to a candidate class that is notcreated as an entity (fails Class Rule 3 after completion of class identification)becomes an operation [2]

3.4.5 Post Processing

Post processing is mainly done to remove unnecessary design elements Postprocessing is done in three areas Trivial relations are converted to operations,trivial associations are converted to attributes and trivial classes are removed Theseare done based on the rules proposed in [2] Since our system, mainly concentrates

on the extraction of Non-Functional Requirements, design of elements has beenadopted from [2,9]

4 Results and Performance Analysis

4.1 Experimentation Setup and Metrics

The input to the system can be a either natural language text document or SRS(Software Requirements Specification) We have taken three standard input systems

as datasets namely ATM, Online Course Registration System and Payroll System.This system has been implemented in Ubuntu environment by employing thelanguages Python and Shell Script for extraction of concepts, attributes and rela-tionships [13] and dot script for visualizing the conceptual model generated [14].Figure2depicts the conceptual model generated and visualized [14] with NFR for

an ATM system requirements specification

In previous model, performance of the functional requirements modelinghas been measured based on three metrics namely Recall, Precision and Over-Specification Equations 1, 2 and 3 depict the formulas for precision, recall andover-specification respectively

Precision ¼ Ncorrect= Nð correctþ NincorrectÞ ð1ÞRecall¼ Ncorrect

Ncorrect specifies the number of design elements correctly identified by the tem and Nincorrectindicates the number of wrong design elements that are identified

sys-as correct Nmissingindicates the number of design elements that are in the answersbut not identified by the system and Nextrashows the number of valid extra classesretrieved by the system

4.2 Performance Evaluation

The three standard input documents taken to evaluate our system are ATM [15],Course Registration system [16] and Payroll System where payroll system is anuser generated input Initially, performance of the NFR classifier has been evaluated

Fig 2 CM-NFR visualization for ATM system

to assure whether the Non-Functional Requirements have been fully extracted bythe system Later, extraction rate of NFR attributes are evaluated using measuressuch as precious, recall and f-measures In addition to that, a comparative analysis

of the identification of NFR attributes between manual and automated process hasbeen evaluated

4.2.1 Evaluation of the Classifier

The performance of the classifier can be measured by calculating the number ofrequirements that are classified correctly versus the number of requirements clas-

sified incorrectly The evaluation was done by taking the requirements documents

of three domains and operating the classifier over it The above process wasrepeated 5 times to ensure that the results are reliable The results obtained aresummarized in the Table1 shown above

4.2.2 Evaluation of NFR Extraction

The performance evaluation of the process of extracting specific Non FunctionalRequirements can be done by calculation precision, recall and F-score based onTrue Positive (TP), False Positive (FP), False Negative (FN) values TP indicatescorrect classification F-score gives a weighted average of the precision and recallvalues The performance was evaluated again by running the process for three casestudies and the results are summarized in the Table2and the Fig.3shown below

Table 1 Performance evaluation of the classi fier

4.2.3 Identifying NFR Attributes

The subjects were given time to analyze NFR attributes and to group themaccording to their category such as cost, interoperability and performance (memory,time, speed) The time taken by the individuals to identify NFR attributes from aPayroll system requirements and automatically generated conceptual model usingthe same payroll system were measured separately and the results are shown inTable3

5 Conclusion and Future Work

This research provides an efficient approach to extract design elements as well astheir Non-Functional Requirements from requirements text document Theextraction and representation of the NFRs also known as soft goals is a novel aspect

of our work through this NFR extraction and visualization in conceptual models,Non-Functional Requirements could be better grasped by the individuals ratherfrom a natural language text In future, this could be extended to more number ofNFRs

Acknowledgments This research is getting supported by UGC, New Delhi, India under Major Research project scheme of Engineering Sciences under the File – F.no 42-129/2013(SR).

Fig 3 Performance evaluation of NFR

Table 3 Time taken by individuals in identifying NFR attributes

conceptual model (s)

1 Meth, H., Brhel, M., Maedche, A.: The state of the art in automated requirements elicitation.

J Inf Softw Technol 55(10), 1695 –1709 (2013) (Elsevier)

2 VidhuBala, R.V., Abirami, S.: Conceptual modeling of explicit natural language functional speci fications J Syst Softw 88(1), 25–41 (2013) (Elsevier)

3 Larman, C.: Applying UML and Patterns —An Introduction to Object-Oriented Analysis and Design and Iterative Development 3rd edn Pearson Education, Ghaziabad (2005)

4 Saleh, K., Al-Zarouni, A.: Capturing non-functional software requirements using the user requirements notation In: Proceedings of The International Research Conference on Innovations in Information Technology, India (2004)

5 Casamayor, A., Godoy, D., Campo, M.: Functional grouping of natural language requirements for assistance in architectural software design J Knowl Based Syst 30(1), 78 –86 (2012) (Elsevier)

6 Deeptimahanti, D.K., Sanyal, R.: Semi-automatic generation of UML models from natural language requirements In: Proceedings of ISEC, ACM, India (2011)

7 Sarkar, S., Sharma, V.S., Agarwal, R.: Creating design from requirements and use cases: bridging the gap between requirement and detailed design In: Proceedings of ISEC, ACM, India (2012)

8 Huertas, C., Reyes, J.R.: NLARE, A natural language processing tool for automatic requirements evaluation In: Proceedings of CUBE, ACM, India (2012)

9 Elbendak, M., Vickers, P., Rossiter, N.: Parsed use case descriptions as a basis for oriented class model generation J Syst Softw 84(7), 1209 –1223 (2011) (Elsevier)

object-10 Bajwa, I.S., Lee, M., Bordbar, B.: Translating natural language constraints to OCL J King Saud Univ Comput Inf Sci 24(2), 117 –128 (2012) (Production and hosting by Elsevier)

11 http://www.nltk.org the of ficial site for NLTK

12 http://nlp.stanford.edu/software/lex-parser.shtmt Parser and POS Tagger

13 Gowsikhaa, D., Abirami, S., Baskaran, R.: Construction of image ontology using low level for image retrivel In: Proceedings of the International Conference on Computer Communication and Informatics, (ICCCI 2012), pp 129 –134 (2012)

14 VidhuBala, R.V., Mala, T., Abirami, S.: Effective visualization of conceptual class diagrams In: Proceedings of International Conference on Recent Advances in Computing and Software Systems pp 1 –6 (2012) doi: 10.1109/RACSS.2012.6212688

15 Rumbaugh, J., Blaha, M., Premerlani, W., Eddy, F., Lorensen,W.: Object-Oriented Modeling and Design Pearson Education, Ghaziabad (1991)

16 Section 1: Course Registration Requirements, IBM Corp, IBM Rational Softwares

A Genetic Algorithm Approach

for Multi-criteria Project Selection

for Analogy-Based Software Cost

Estimation

Sweta Kumari and Shashank Pushkar

Abstract This paper presents genetic algorithms as multi-criteria project selectionfor improving the Analogy Based Estimation (ABE) process, which is suitable toreuse past project experience to create estimation of the new projects An attempthas also been made to create a multi-criteria project selection problem with andwithout allowing for interactive effects between projects based on criteria which aredetermined by the decision makers Two categories of projects are also presentedfor comparison purposes with other traditional optimization methods and theexperimented results show the capability of the proposed Genetic Algorithm basedmethod in multi-criteria project selection problem and it can be used as an efficientsolution to the problem that will enhance the ABE process Here, Mean AbsoluteRelative Error (MARE) is used to evaluate the performance of ABE process and ithas been found that interactive effects between projects may change the results

Keywords Software cost estimation
Analogy based estimation
Geneticalgorithm
Multi-criteria decision making
Nonlinear integer programming

1 Introduction

Success of software organizations rely on proper management activities such asplanning, budgeting, scheduling, resource allocation and effort requirements forsoftware projects Software effort estimation is the process of making an approximatejudgment of the costs of software Inaccurate estimation of software cost lowers theproficiency of the project, wastes the company’s budget and can result in failure of the

S Kumari ( &)
S Pushkar

Department of CSE, BIT, Mesra, Ranchi, India

e-mail: swetak44@gmail.com

S Pushkar

e-mail: shashank_pushkar@yahoo.com

© Springer India 2015

L.C Jain et al (eds.), Computational Intelligence in Data Mining - Volume 3,

Smart Innovation, Systems and Technologies 33, DOI 10.1007/978-81-322-2202-6_2

13

entire project Broadly, there are two types of cost estimation methods: Algorithmicand Non-algorithmic Algorithmic method calculates cost of the software by usingformula or some experimental equations whereas Non-algorithmic methods use ahistorical data that is related to the previously completed similar projects for calcu-lating the cost of the software [1] Analogy based estimation is a Non-algorithmicmethod in that a critical role is played by the similarity measures between a pair ofprojects Here, a distance is calculated between the software project being estimatedand each of the historical software projects It thenfinds the most similar project that

is used to estimate the cost Estimation by analogy is essentially a form of Case BasedReasoning [2] However, as it is argued in [3] there are certain advantages in respectwith rule based systems, such as the fact that users are keen to accept solutions fromanalogy based techniques, rather than solutions derived from uncomfortable chains

of rules or neural nets Naturally, there are some difficulties with analogy-basedestimation such as lack of appropriate analogues and issues with selecting and usingthem Choosing an appropriate set of projects participating in cost estimation processare very important for any organizations to achieve their goals In this process, severalreasons involve such as the number of investment projects, presence of multipledecision criteria such as takings maximization or risk minimization, business andoperational rules such as budget limits and time windows for starting dates Projectselection is a difficult task, if the project interactions in terms of multiple selectioncriteria and information of preferences of decision-makers are taken into account,mainly in the presence of a huge number of projects

2 Related Works

Various methods and mathematical models have been developed to deal with theproblem of selecting and scheduling projects Since the pioneering ranking methodfrom [4] other methods have been proposed: Scoring [5], Analytical HierarchyProcess [6,7] and Goal Programming [8] among others However, these methodsthink that project interdependencies do not exist [9] Other authors have proposedproject selection models that deal with the existence of interdependencies.According to [8], these models, classified by the fundamental solution method are:Dynamic Programming [10] models reflect interdependency in special cases;Quadratic/Linear 0–1 programming models have a quadratic objective function andlinear constraints, they limit interdependency only in the objective and between twoprojects Quadratic/Quadratic 0–1 programming [11] with interdependency betweentwo projects in the objective function and in the resource constraints; and Nonlinear

0–1 programming, with interdependency reflected in the objective function and theconstraints among as many projects as necessary [9, 12] presents a linear 0–1programming model for the selection and scheduling of projects, that includestechnical interdependency Medaglia et al [13] also present some features of