Research and development in intelligent systems XXXIII

Max Bramer Miltos Petridis Editors Research and Development in Intelligent Systems XXXIII Incorporating Applications and Innovations in Intelligent Systems XXIV Proceedings of AI-2016, The Thirty-Sixth SGAI International Conference on Innovative Techniques and Applications of Artificial Intelligence Research and Development in Intelligent Systems XXXIII Incorporating Applications and Innovations in Intelligent Systems XXIV Max Bramer ⋅ Miltos Petridis Editors Research and Development in Intelligent Systems XXXIII Incorporating Applications and Innovations in Intelligent Systems XXIV Proceedings of AI-2016, The Thirty-Sixth SGAI International Conference on Innovative Techniques and Applications of Artificial Intelligence 123 Editors Max Bramer School of Computing University of Portsmouth Portsmouth UK ISBN 978-3-319-47174-7 DOI 10.1007/978-3-319-47175-4 Miltos Petridis School of Computing, Engineering and Mathematics University of Brighton Brighton UK ISBN 978-3-319-47175-4 (eBook) Library of Congress Control Number: 2016954594 © Springer International Publishing AG 2016 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 This Springer imprint is published by Springer Nature The registered company is Springer International Publishing AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland Programme Chairs’ Introduction This volume comprises the refereed papers presented at AI-2016, the Thirty-sixth SGAI International Conference on Innovative Techniques and Applications of Artificial Intelligence, held in Cambridge in December 2016 in both the technical and the application streams The conference was organised by SGAI, the British Computer Society Specialist Group on Artificial Intelligence The technical papers included present new and innovative developments in the field, divided into sections on Knowledge Discovery and Data Mining, Sentiment Analysis and Recommendation, Machine Learning, AI Techniques, and Natural Language Processing This year’s Donald Michie Memorial Award for the best-refereed technical paper was won by a paper entitled “Harnessing Background Knowledge for E-learning Recommendation” by B Mbipom, S Craw and S Massie (Robert Gordon University, Aberdeen, UK) The application papers included present innovative applications of AI techniques in a number of subject domains This year, the papers are divided into sections on legal liability, medicine and finance, telecoms and e-Learning, and genetic algorithms in action This year’s Rob Milne Memorial Award for the best-refereed application paper was won by a paper entitled “A Genetic Algorithm Based Approach for the Simultaneous Optimisation of Workforce Skill Sets and Team Allocation” by A.J Starkey and H Hagras (University of Essex, UK), S Shakya and G Owusu (British Telecom, UK) The volume also includes the text of short papers presented as posters at the conference On behalf of the conference organising committee, we would like to thank all those who contributed to the organisation of this year’s programme, in particular the programme committee members, the executive programme committees and our administrators Mandy Bauer and Bryony Bramer Max Bramer, Technical Programme Chair, AI-2016 Miltos Petridis, Application Programme Chair, AI-2016 v Acknowledgements/Committees AI-2016 Conference Committee Prof Max Bramer, University of Portsmouth (Conference Chair) Prof Max Bramer, University of Portsmouth (Technical Programme Chair) Prof Miltos Petridis, University of Brighton (Application Programme Chair) Dr Jixin Ma, University of Greenwich (Deputy Application Programme Chair) Prof Adrian Hopgood, University of Liege, Belgium (Workshop Organiser) Rosemary Gilligan (Treasurer) Dr Nirmalie Wiratunga, Robert Gordon University, Aberdeen (Poster Session Organiser) Andrew Lea, Primary Key Associates Ltd (AI Open Mic and Panel Session Organiser) Dr Frederic Stahl, University of Reading (Publicity Organiser) Dr Giovanna Martinez, Nottingham Trent University and Christo Fogelberg Palantir Technologies (FAIRS 2016) Prof Miltos Petridis, University of Brighton and Prof Thomas Roth-Berghofer University of West London (UK CBR Organisers) Mandy Bauer, BCS (Conference Administrator) Bryony Bramer, (Paper Administrator) Technical Executive Programme Committee Prof Max Bramer, University of Portsmouth (Chair) Prof Frans Coenen, University of Liverpool Dr John Kingston, University of Brighton Prof Dan Neagu, University of Bradford Prof Thomas Roth-Berghofer, University of West London Dr Nirmalie Wiratunga, Robert Gordon University, Aberdeen vii viii Acknowledgements/Committees Applications Executive Programme Committee Prof Miltos Petridis, University of Brighton (Chair) Mr Richard Ellis, Helyx SIS Ltd Ms Rosemary Gilligan, University of Hertfordshire Dr Jixin Ma, University of Greenwich (Vice-Chair) Dr Richard Wheeler, University of Edinburgh Technical Programme Committee Andreas Albrecht (Middlesex University) Abdallah Arioua (IATE INRA France) Raed Batbooti (University of Swansea UK (PhD Student), University of Basra (Lecturer)) Lluís Belanche (Universitat Politecnica de Catalunya, Barcelona, Catalonia, Spain) Yaxin Bi (Ulster University, UK) Mirko Boettcher (University of Magdeburg; Germany) Max Bramer (University of Portsmouth) Krysia Broda (Imperial College; University of London) Ken Brown (University College Cork) Charlene Cassar (De Montfort University UK) Frans Coenen (University of Liverpool) Ireneusz Czarnowski (Gdynia Maritime University; Poland) Nicolas Durand (Aix-Marseille University) Frank Eichinger (CTS EVENTIM AG & Co KGaA, Hamburg, Germany) Mohamed Gaber (Robert Gordon University, Aberdeen, UK) Hossein Ghodrati Noushahr (De Montfort University, Leicester, UK) Wael Hamdan (MIMOS Berhad., Kuala Lumpur, Malaysia) Peter Hampton (Ulster University, UK) Nadim Haque (Capgemini) Chris Headleand (University of Lincoln, UK) Arjen Hommersom (Open University, The Netherlands) Adrian Hopgood (University of Liège, Belgium) John Kingston (University of Brighton) Carmen Klaussner (Trinity College Dublin Ireland) Ivan Koychev (University of Sofia) Thien Le (University of Reading) Nicole Lee (University of Hong Kong) Anne Liret (British Telecom France) Fernando Lopes (LNEG-National Research Institute; Portugal) Stephen Matthews (Newcastle University) Silja Meyer-Nieberg (Universitat der Bundeswehr Munchen Germany) Acknowledgements/Committees Roberto Micalizio (Universita’ di Torino) Daniel Neagu (University of Bradford) Lars Nolle (Jade University of Applied Sciences; Germany) Joanna Isabelle Olszewska (University of Gloucestershire UK) Dan O’Leary (University of Southern California) Juan Jose Rodriguez (University of Burgos) Thomas Roth-Berghofer (University of West London) Fernando Saenz-Perez (Universidad Complutense de Madrid) Miguel A Salido (Universidad Politecnica de Valencia) Rainer Schmidt (University Medicine of Rostock; Germany) Frederic Stahl (University of Reading) Simon Thompson (BT Innovate) Jon Timmis (University of York) M.R.C van Dongen (University College Cork) Martin Wheatman (Yagadi Ltd.) Graham Winstanley (University of Brighton) Nirmalie Wiratunga (Robert Gordon University) Application Programme Committee Hatem Ahriz (Robert Gordon University) Tony Allen (Nottingham Trent University) Ines Arana (Robert Gordon University) Mercedes Arguello Casteleiro (University of Manchester) Ken Brown (University College Cork) Sarah Jane Delany (Dublin Institute of Technology) Richard Ellis (Helyx SIS Ltd.) Roger Evans (University of Brighton) Andrew Fish (University of Brighton) Rosemary Gilligan (University of Hertfordshire) John Gordon (AKRI Ltd.) Chris Hinde (Loughborough University) Adrian Hopgood (University of Liege, Belgium) Stelios Kapetanakis (University of Brghton) Alice Kerly Jixin Ma (University of Greenwich) Lars Nolle (Jade University of Applied Sciences) Miltos Petridis (University of Brighton) Miguel A Salido (Universidad Politecnica de Valencia) Roger Tait (University of Cambridge) Richard Wheeler (Edinburgh Scientific) ix Contents Research and Development in Intelligent Systems XXXIII Best Technical Paper Harnessing Background Knowledge for E-Learning Recommendation Blessing Mbipom, Susan Craw and Stewart Massie Knowledge Discovery and Data Mining Category-Driven Association Rule Mining Zina M Ibrahim, Honghan Wu, Robbie Mallah and Richard J.B Dobson 21 A Comparative Study of SAT-Based Itemsets Mining Imen Ouled Dlala, Said Jabbour, Lakhdar Sais and Boutheina Ben Yaghlane 37 Mining Frequent Movement Patterns in Large Networks: A Parallel Approach Using Shapes Mohammed Al-Zeyadi, Frans Coenen and Alexei Lisitsa 53 Sentiment Analysis and Recommendation Emotion-Corpus Guided Lexicons for Sentiment Analysis on Twitter Anil Bandhakavi, Nirmalie Wiratunga, Stewart Massie and P Deepak Context-Aware Sentiment Detection from Ratings Yichao Lu, Ruihai Dong and Barry Smyth 71 87 Recommending with Higher-Order Factorization Machines 103 Julian Knoll xi xii Contents Machine Learning Multitask Learning for Text Classification with Deep Neural Networks 119 Hossein Ghodrati Noushahr and Samad Ahmadi An Investigation on Online Versus Batch Learning in Predicting User Behaviour 135 Nikolay Burlutskiy, Miltos Petridis, Andrew Fish, Alexey Chernov and Nour Ali A Neural Network Test of the Expert Attractor Hypothesis: Chaos Theory Accounts for Individual Variance in Learning 151 P Chassy AI Techniques A Fast Algorithm to Estimate the Square Root of Probability Density Function 165 Xia Hong and Junbin Gao 3Dana: Path Planning on 3D Surfaces 177 Pablo Moz, María D R-Moreno and Bonifacio Castaño Natural Language Processing Covert Implementations of the Turing Test: A More Level Playing Field? 195 D.J.H Burden, M Savin-Baden and R Bhakta Context-Dependent Pattern Simplification by Extracting Context-Free Floating Qualifiers 209 M.J Wheatman Short Papers Experiments with High Performance Genetic Programming for Classification Problems 221 Darren M Chitty Towards Expressive Modular Rule Induction for Numerical Attributes 229 Manal Almutairi, Frederic Stahl, Mathew Jennings, Thien Le and Max Bramer OPEN: New Path-Planning Algorithm for Real-World Complex Environment 237 J.I Olszewska and J Toman 382 L Hernandez Mengesha and C.J James-Reynolds Conclusions This work explored the use of an iGA hardware controller as the interface for producing evolutionary art We were particularly interested in a user perspective and issues in adapting the controller User engagement varied considerably and depended on their interest, but the perceived lack of visible change when sliders were adjusted led to a feeling of not being in control Users seemed more able to successfully complete goal oriented tasks when they understood the concept behind the interface References James-Reynolds, C., Currie, E.: Eugene: A Generic Interactive Genetic Algorithm Research and Development in Intelligent Systems XXXII, pp 361–366 Springer International Publishing, London (2015) Fry, B., Reas, C.: The Processing Language https://processing.org/ Accessed 20 Aug 2016 Galanter, P.: What is generative art? Complexity theory as a context for art theory In: GA2003— 6th Generative Art Conference (2003) Verostko, R.: Notes on algorithmic art in the ISEA’94 Catalogue In: The Fifth International Symposium on Electronic Art, p 61 University of Art & Design, Helsinki, Finland (1994) Mandelbrot, B.: Fractal aspects of the iteration of z → λ z (1-z) for complex λ and z Non-Linear Dynamics Ann N Y Acad Sci 357, 249–259 (1979) Julia, G.: Mémoire sur l’iteration des fonctions rationnelles Journal de Mathématiques Pures et Appliquées 8, 47–245 (1918) Barnsley, M.: Fractals Everywhere Academic Press, Boston (1988) Turing, A.M.: Computing Machinery and Intelligence Mind 49, 433–460 (1950) Holland, J.H.: Adaptation in Natural and Artificial Systems University of Michigan Press, Ann Arbor (1975) 10 Mitchell, M.: An Introduction To Genetic Algorithms Massachusetts London, England, Cambridge (2002) (Fifth printing) 11 Bremermann, H.J.: Optimization through evolution and recombination Self-organizing Syst 93, 106 (1962) 12 Dianati, M., Song, I., Treiber, M.: An introduction to genetic algorithms and evolution strategies Technical report University of Waterloo, Ontario, Canada (2002) 13 Taylor, T., Dorin, A., Korb, K.: Digital genesis: computers, evolution and artificial life In: Presented at the 7th Munich-Sydney-Tilburg Philosophy of Science Conference: Evolutionary Thinking, University of Sydney, 20–22 (2014) 14 The MIDI 1.0 Specification https://www.midi.org/specifications Accessed 20 Aug 2016 15 De Jong K.: Analysis of the Behavior of a Class of Genetic Adaptive Systems Technical Report no 185 University of Michigan (1975) Incorporating Emotion and Personality-Based Analysis in User-Centered Modelling Mohamed Mostafa, Tom Crick, Ana C Calderon and Giles Oatley Abstract Understanding complex user behaviour under various conditions, scenarios and journeys is fundamental to improving the user-experience for a given system Predictive models of user reactions, responses—and in particular, emotions—can aid in the design of more intuitive and usable systems Building on this theme, the preliminary research presented in this paper correlates events and interactions in an online social network against user behaviour, focusing on personality traits Emotional context and tone is analysed and modelled based on varying types of sentiments that users express in their language using the IBM Watson Developer Cloud tools The data collected in this study thus provides further evidence towards supporting the hypothesis that analysing and modelling emotions, sentiments and personality traits provides valuable insight into improving the user experience of complex social computer systems Keywords Emotions · Personality · Sentiment analysis · User experience · Social networking · Affective computing The original extended version of this paper is available here: http://arxiv.org/abs/1608.03061 M Mostafa · T Crick (B) · A.C Calderon Department of Computing & Information Systems, Cardiff Metropolitan University, Cardiff, UK e-mail: tcrick@cardiffmet.ac.uk M Mostafa e-mail: momostafa@cardiffmet.ac.uk A.C Calderon e-mail: acalderon@cardiffmet.ac.uk G Oatley School of Engineering & Information Technology, Murdoch University, Murdoch, Australia e-mail: g.oatley@murdoch.edu.au © Springer International Publishing AG 2016 M Bramer and M Petridis (eds.), Research and Development in Intelligent Systems XXXIII, DOI 10.1007/978-3-319-47175-4_29 383 384 M Mostafa et al Introduction As computer systems and applications have become more widespread and complex, with increasing demands and expectations of ever-more intuitive human-computer interactions, research in modelling, understanding and predicting user behaviour demands has become a priority across a number of domains In these application domains, it is useful to obtain knowledge about user profiles or models of software applications, including intelligent agents, adaptive systems, intelligent tutoring systems, recommender systems, e-commerce applications and knowledge management systems [12] Furthermore, understanding user behaviour during system events leads to a better informed predictive model capability, allowing the construction of more intuitive interfaces and an improved user experience This work can be applied across a range of socio-technical systems, impacting upon both personal and business computing We are particularly interested in the relationship between digital footprint and behaviour and personality [6, 8] A wide range of pervasive and often publicly available datasets encompassing digital footprints, such as social media activity, can be used to infer personality [4, 9] and development of robust models capable of describing individuals and societies [5] Social media has been used in varying computer system approaches; in the past this has mainly been the textual information contained in blogs, status posts and photo comments [1, 2], but there is also a wealth of information in the other ways of interacting with online artefacts From sharing and gathering of information and data, to catering for marketing and business needs; it is now widely used as technical support for computer system platforms The work presented in this paper is builds upon previous work in psycholinguistic science and aims to provide further insight into how the words and constructs we use in our daily life and online interactions reflect our personalities and our underlying emotions As part of this active research field, it is widely accepted that written text reflects more than the words and syntactic constructs, but also conveys emotion and personality traits [10] As part of our work, the IBM Watson Tone Analyzer (part of the IBM Watson Developer Cloud toolchain) has been used to identify emotion tones in the textual interactions in an online system, building on previous work in this area that shows a strong correlation between the word choice and personality, emotions, attitude and cognitive processes, providing further evidence that it is possible to profile and potentially predict users identity [3] The Linguistic Inquiry and Word Count (LIWC) psycholinguistics dictionary [11, 13] is used to find psychologically meaningful word categories from word usage in writing; the work presented here provides a modelling and analysis framework, as well as associated toolchain, for further application to larger datasets to support the research goal of improving usercentered modelling The rest of the paper is structured as follows: in Sects and we present our data, the statistical analysis and identify the key elements of our model; in Sect we summarise the main contributions of this paper, as well as making clear recommendations for future research Incorporating Emotion and Personality-Based Analysis … 385 Data Analysis and Feature Extraction Our dataset comes from an online portal for a European Union international scholarship mobility hosted at a UK university The dataset was generated from interactions between users and a complex online information system, namely the online portal for submitting applications The whole dataset consists of users (N = 391), interactions and comments (N = 1390) as responses to system status and reporting their experience with using the system Google Analytics has been used to track user behaviour and web statistics (such as impressions); this data from has been used to identify the server’s status and categorised the status as two stages: Idle, where the system had a higher number of active sessions; and marked as Failure, where the system had a lower number of sessions engaged Interactions were first grouped by server status, then sent to the IBM Watson Tone Analyzer to generate emotion social tone scores In what follows Failure status shows a significant difference in overall Anger in different status; furthermore, the Joy parameter shows a significant difference with the system in Idle and Failure status However Fear and Sadness parameters is almost the same, even with the system in Idle status We identified the user’s personality based on analysis of their Facebook interactions, namely by collecting all comments from the users, again using the IBM Watson Personality Insights tool However, a number of users in the dataset had completed the Big Five questionnaire (N = 44); for these users, their Big Five scores have been used instead The second stage involved grouping the comments based on server status and segmenting these interactions by user; this allowed us to investigate the impact of server status in the emotion of the user and investigate the Big Five dimension as a constant parameter By investigating the relationship between personality trait dimensions and the social emotion tones, we are able to find the highest correlation to identify the key elements of the potential model by applying linear regression and Pearson correlation This allowed building of a neural network multilayer perception using the potential key elements with higher correlations The data collected from the social media interactions was grouped by users and via IBM Watson Personality Insights, we were able to identify the Big Five personality traits for each user Using the IBM Watson Tone Analyzer, the data was grouped by user’s comments and server status (Failure, Idle) to identify the social emotion tone for each user Statistical Analysis and Key Elements of Model As part of modelling the users’ responses and behaviour, in order to build a conceptual framework model, we applied linear regression to investigate the relationship between the Big Five personality dimensions and the emotion tones Linear regressions (presented in Table and Fig 1) not show significant correlations between the Big Five dimensions and the social emotion tones There are, Extraversion B t 0.162 1.642 0.064 0.831 0.114 1.142 0.172 1.549 0.446 4.179 −0.185 −1.906 Openness B t Sig (Constant) 0.356 3.282 0.001 Anger −0.063 −0.735 0.463 Disgust 0.478 4.354 Fear 0.065 0.534 0.594 Joy 0.066 0.561 0.575 Sadness −0.226 −2.118 0.035 Table Linear regression coefficients Sig 0.101 0.406 0.253 0.122 0.057 B t 0.16 1.623 0.124 1.592 0.255 2.551 0.04 0.356 0.436 4.058 −0.03 −0.313 Sig 0.105 0.112 0.011 0.722 0.754 Conscientiousness B t Sig 0.297 2.831 0.005 0.024 0.293 0.769 −0.061 −0.574 0.566 0.093 0.783 0.434 0.188 1.652 0.099 0.014 0.132 0.895 Agreeableness B t Sig 0.828 9.934 0.116 1.767 0.078 −0.363 −4.303 −0.023 −0.241 0.81 −0.487 −5.39 0.233 2.841 0.005 Neuroticism 386 M Mostafa et al Incorporating Emotion and Personality-Based Analysis … 387 Fig Scatterplots of the Big Five dimension (dependent variables) and social emotion tones (independent variables) however correlations that can be used as key elements for the model; namely the correlation of Openness and Disgust (0.479), Extraversion and Joy (0.446 with p-value of 0), Conscientiousness and Joy (0.436),and Disgust with 0.255.Agreeableness, does not appear to have a high impact in the social emotion parameters, with the highest correlation being 0.188 with Joy, which can be overlooked as a useful factor in the model Neuroticism and Disgust is −0.363, Joy is −0.487 and p-value is zero is both cases; and Sadness with 0.233 All correlation values are 0.3 and