Studies in Computational Intelligence 442 Editor-in-Chief Prof.JanuszKacprzyk SystemsResearchInstitute PolishAcademyofSciences ul.Newelska6 01-447Warsaw Poland E-mail:[email protected] Forfurthervolumes: http://www.springer.com/series/7092 Ivan Jordanov and Lakhmi C. Jain (Eds.) Innovations in Intelligent Machines -3 Contemporary Achievements in Intelligent Systems ABC Editors Dr.IvanJordanov Prof.LakhmiC.Jain UniversityofPortsmouth SchoolofElectricalandInformation UK Engineering UniversityofSouthAustralia Adelaide SouthAustralia Australia ISSN1860-949X e-ISSN1860-9503 ISBN978-3-642-32176-4 e-ISBN978-3-642-32177-1 DOI10.1007/978-3-642-32177-1 SpringerHeidelbergNewYorkDordrechtLondon LibraryofCongressControlNumber:2012943844 (cid:2)c Springer-VerlagBerlinHeidelberg2013 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpartofthe materialisconcerned,specificallytherightsoftranslation,reprinting,reuseofillustrations,recitation,broad- casting,reproductiononmicrofilmsorinanyotherphysicalway,andtransmissionorinformationstorage andretrieval,electronicadaptation,computersoftware,orbysimilarordissimilarmethodologynowknown orhereafterdeveloped.Exemptedfromthislegalreservationarebriefexcerptsinconnectionwithreviews orscholarly analysis ormaterial suppliedspecifically forthepurposeofbeingentered andexecuted ona computersystem,forexclusive usebythepurchaser ofthework.Duplication ofthis publication orparts thereofispermittedonlyundertheprovisionsoftheCopyrightLawofthePublisher’slocation,initscur- rentversion,andpermissionforusemustalways beobtained fromSpringer. Permissionsforusemaybe obtainedthroughRightsLinkattheCopyrightClearanceCenter.Violationsareliabletoprosecutionunder therespectiveCopyrightLaw. Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant protectivelawsandregulationsandthereforefreeforgeneraluse. Whiletheadviceandinformationinthisbookarebelievedtobetrueandaccurateatthedateofpublication, neither the authors northe editors nor the publisher can accept any legal responsibility for any errors or omissionsthatmaybemade.Thepublishermakesnowarranty,expressorimplied,withrespecttothematerial containedherein. Printedonacid-freepaper SpringerispartofSpringerScience+BusinessMedia(www.springer.com) Preface This research volume is a continuation of our previous volumes on intelligent machines. We laid the foundation of intelligent machines in SCI Series Volume 70 by including the possible and successful applications of computational intelligence paradigms in machines for mimicking the human behaviour. The present volume includes the recent advances in intelligent paradigms and innovative applications such as document processing, language translation, English academic writing, crawling system for web pages, web-page retrieval technique, aggregate k-Nearest Neighbour for answering queries, context-aware guide, recommendation system for museum, meta-learning environment, case-based reasoning approach for adaptive modelling in exploratory learning, discussion support system for understanding research papers, system for recommending e-Learning courses, community site for supporting multiple motor-skill development, community size estimation of internet forum, lightweight reprogramming for wireless sensor networks, adaptive traffic signal controller and virtual disaster simulation system. Modern information technology relies on intelligent systems that can learn and reason over a range of knowledge hidden in available datasets. Achieving these objectives involve applying machine learning biology and nature inspired methods to inductively construct computational and mathematical models (that can use explicit or implicit human supervision) and gain insight in terms of patterns and relationships hidden into the datasets in hand. This ‘insight’ helps the intelligent systems to reason and learn and to use the extracted knowledge for prediction of trends and tendencies, for processes and products monitoring and control, for fault detection and diagnosing in a wide range of application areas. The aim of this edited book is to promote current theoretical and application oriented Intelligent systems research (more specifically in the field of neural networks computing) and to present examples of experimental and real-world investigations that demonstrate contemporary achievements and advances in the area. Leading researchers contribute articles presenting works from this multi-faceted and burgeoning area of research at both theoretical and application levels, covering a variety of topics related to intelligent systems. This book is directed to engineers, scientists, researchers, professor and the undergraduate/postgraduate students who wish to explore the applications of intelligent paradigms further. We are grateful to the authors and reviewers for their contribution and appreciate the assistance provided by the editorial team of Springer-Verlag. Dr. Ivan Jordanov, University of Portsmouth, UK Professor Lakhmi C. Jain, University of South Australia, Adelaide, Australia Books on Innovations in Intelligent Machines • Chahl, J.S., Jain, L.C., Mizutani, A. and Sato-Ilic, M., Innovations in Intelligent Machines 1, Springer-Verlag, Germany, 2007. • Watanabe, T. and Jain, L.C., Innovations in Intelligent Machines 2: Intelligent Paradigms and Applications, Springer-Verlag, Germany, 2012. • Jordanov, I. and Jain, L.C., Innovations in Intelligent Machines 3: Contemporary Achievements in Intelligent Systems, Springer-Verlag, Germany, 2012. Contents 1 An Introduction to Contemporary Achievements in Intelligent Systems ......... 1 Jeffrey W. Tweedale, Ivan Jordanov 1 Introduction ........................................................................................................ 1 1.1 What Is a System? ...................................................................................... 2 1.2 What Is Intelligence? .................................................................................. 3 1.3 What Is an Intelligent System? ................................................................... 4 1.4 What Is AI? ................................................................................................. 4 1.5 Putting AI to Work ..................................................................................... 5 1.6 New AI ....................................................................................................... 6 1.7 Intelligent Paradigms .................................................................................. 6 1.8 Knowledge .................................................................................................. 7 1.9 Other Effects ............................................................................................... 8 1.10 Why Agents? ............................................................................................ 9 2 Chapters Included in the Book ........................................................................... 9 3 Conclusion ........................................................................................................ 12 References ............................................................................................................. 13 2 Market Power Assessment Using Hybrid Fuzzy Neural Network ...................15 Kirti Pal, Manjaree Pandit, Laxmi Srivastava 1 Nomenclature ................................................................................................... 15 2 Introduction ...................................................................................................... 16 3 Market Power ................................................................................................... 17 4 FNN Approach for Open Electricity Market .................................................... 19 5 Hybrid Fuzzy Neural Network Based Power Market Assessment ................... 23 6 Training and Testing Detail .............................................................................. 27 7 Conclusion ........................................................................................................ 34 References ............................................................................................................. 35 3 Coaching Robots: Online Behavior Learning from Human Subjective Feedback................................................................................................................37 Masakazu Hirkoawa, Kenji Suzuki 1 Introduction ...................................................................................................... 37 2 Methodology .................................................................................................... 39 2.1 Coaching ................................................................................................... 39 2.2 Reinforcement Learning in Continuous State-Action Space .................... 42 2.3 Implementation of Coaching .................................................................... 42 2.4 The Basic Problem and the Solution Approach ........................................ 43 3 Reward Function and Its Learning Efficiency .................................................. 44 VIII Contents 4 Evaluation of Behavior Learning ..................................................................... 47 4.1 Characteristics of Human Evaluation ....................................................... 47 4.2 Experiment with a Simulated Robotic Agent ........................................... 48 4.3 Results ...................................................................................................... 48 4.4 Experiment with a Real Robotic Agent .................................................... 50 5 Conclusions ...................................................................................................... 50 References ............................................................................................................. 51 4 Persian Vowel Recognition Using the Combination of Haar Wavelet and Neural Network ..............................................................................................53 Mohammad Mehdi Hosseini, Abdorreza Alavi Gharahbagh 1 Introduction ...................................................................................................... 53 2 Proposed Method .............................................................................................. 54 2.1 Lip Localization ........................................................................................ 54 2.1.1 Face Detection ............................................................................... 55 2.1.2 Lip Region of Interest .................................................................... 55 2.2 Lip Segmentation ...................................................................................... 56 2.2.1 Pre Processing ............................................................................... 56 2.2.2 Color Transform ............................................................................ 56 2.2.3 Wavelet .......................................................................................... 58 2.2.4 Lip Segmentation ........................................................................... 58 2.2.5 Morphological Filtering ................................................................. 59 2.2.6 Post Processing .............................................................................. 59 3 Vowel Recognition ........................................................................................... 59 3.1 Feature Vector .......................................................................................... 59 3.2 Neural Network ........................................................................................ 60 4 Material and Methods ....................................................................................... 62 5 Results .............................................................................................................. 65 6 Conclusion ........................................................................................................ 67 References ............................................................................................................. 67 5 The Reproduction of the Physiological Behaviour of the Axon of Nervous Cells by Means of Finite Element Models ...........................................................69 Simona Elia, Patrizia Lamberti Introduction ........................................................................................................... 69 1 Proposed FEM Axon Models ........................................................................... 71 1.1 Model A – Axon with Membrane Domain ............................................... 73 1.2 Model B – Axon with Thin Layer Approximation ................................... 76 1.3 Model A vs. Model B ............................................................................... 77 2 AP Temperature Dependence and Feature of a FEM Approach ...................... 80 3 Best Numerical Model with Respect to the Objective of the Analysis .................................................................................................. 82 3.1 Temperature Dependence of the Firing Effect .......................................... 82 3.2 The Propagation of the AP along the Membrane Domain ........................ 84 Conclusions and Future Work ............................................................................... 85 References ............................................................................................................. 86 Contents IX 6 A Study of a Single Multiplicative Neuron (SMN) Model for Software Reliability Prediction ............................................................................................89 S. Chatterjee, J.B. Singh, S. Nigam, L.N. Upadhyaya 1 Introduction ...................................................................................................... 89 2 The Single Multiplicative Neuron Model ......................................................... 91 3 Learning Rule for the Single Multiplicative Neuron Model ............................. 91 3.1 Back Propagation (BP) Learning Algorithm ............................................ 91 3.2 The Genetic Algorithm ............................................................................. 92 4 Results and Discussion ..................................................................................... 94 5 Comparison ...................................................................................................... 96 6 Concluding Remark ........................................................................................ 100 References ........................................................................................................... 100 7 Numerical Treatment for Painlevé Equation I Using Neural Networks and Stochastic Solvers .........................................................................................103 Muhammad Asif Zahoor Raja, Junaid Ali Khan, Siraj-ul-Islam Ahmad, Ijaz Mansoor Qureshi 1 Introduction .................................................................................................... 103 2 Neural Network Mathematical Model ............................................................ 104 3 Stochastic Solvers ........................................................................................... 106 4 Simulation and Results ................................................................................... 110 5 Conclusion ...................................................................................................... 116 References ........................................................................................................... 116 8 An Investigation into the Adaptive Capacity of Recurrent Neural Networks ..............................................................................................................119 N.H. Siddique, B.P. Amavasai 1 Introduction .................................................................................................... 119 2 Overview of Recurrent Architectures ............................................................. 120 3 Proposed Network Architecture ..................................................................... 122 4 Training Algorithm of the RNN ..................................................................... 128 5 Experimentation and Analysis of Results ....................................................... 129 6 Conclusion ...................................................................................................... 135 References ........................................................................................................... 136 9 An Extended Approach of a Two-Stage Evolutionary Algorithm in Artificial Neural Networks for Multiclassification Tasks ................................139 Antonio J. Tallón-Ballesteros, César Hervás-Martínez, Pedro A. Gutiérrez 1 Introduction .................................................................................................... 139 2 Methodology .................................................................................................. 140 2.1 Evolutionary Artificial Neural Networks Based on Sigmoidal and Product Units .................................................................................... 140 2.2 Two-Stage Evolutionary Algorithm ....................................................... 141 3 Proposal Description ...................................................................................... 143 X Contents 4 Experimentation ............................................................................................. 144 4.1 Data Sets and Validation Technique ....................................................... 144 4.2 Common Parameters of the Different Methodologies and Specific Parameters Depending on the Dataset .................................................... 145 4.3 Nonparametric Statistical Analysis ......................................................... 147 5 Results ............................................................................................................ 147 5.1 Results Applying TSEA, EDDSig and TSEASig ................................... 147 5.1.1 Statistical Analysis ...................................................................... 148 5.1.2 Analysis of Computational Cost .................................................. 149 5.2 Results Obtained with a Good Number of Classifiers ............................ 150 6 Conclusions .................................................................................................... 151 References ........................................................................................................... 152 Author Index ............................................................................................................155