Table Of Content

Computational Intelligence Techniques for Trading and Investment Computational intelligence, a sub- branch of artificial intelligence, is a field which draws on the natural world and adaptive mechanisms in order to study behaviour in changing complex environments. This book provides an interdisci- plinary view of current technological advances and challenges concerning the application of computational intelligence techniques to financial time-s eries forecasting, trading and investment. The book is divided into five parts. The first part introduces the most important computational intelligence and financial trading concepts, while also presenting the most important methodologies from these different domains. The second part is devoted to the application of traditional computational intelligence techniques to the fields of financial forecasting and trading, and the third part explores the applications of artificial neural networks in these domains. The fourth part delves into novel evolutionary-b ased hybrid methodologies for trading and portfolio management, while the fifth part presents the applications of advanced computational intelligence modelling techniques in financial forecasting and trading. This volume will be useful for graduate and postgraduate students of finance, computational finance, financial engineering and computer science. Practition- ers, traders and financial analysts will also benefit from this book. Christian Dunis is Emeritus Professor of Banking and Finance at Liverpool John Moores University, UK and Joint General Manager of global risk and new products at Horus Partners Wealth Management Group SA, Switzerland. Spiros Likothanassis is Professor and Director at the Pattern Recognition Labor atory in the Department of Computer Engineering and Informatics at the University of Patras, Greece. Andreas Karathanasopoulos is Senior Lecturer in Finance and Risk Manage- ment at the University of East London, UK. Georgios Sermpinis is Senior Lecturer in Economics at the University of Glasgow, UK. Konstantinos Theofilatos is a Post-Doctoral Researcher in the Department of Computer Engineering and Informatics at the University of Patras, Greece. Routledge advances in experimental and computable economics Edited by K. Vela Velupillai and Stefano Zambelli University of Trento, Italy 1 The Economics of Search Brian and John McCall 2 Classical Econophysics Paul Cockshott, Allin F. Cottrell, Gregory John Michaelson, Ian P. Wright and Victor Yakovenko 3 The Social Epistemology of Experimental Economics Ana Cordeiro dos Santos 4 Computable Foundations for Economics K. Vela Velupillai 5 Neuroscience and the Economics of Decision Making Edited by Alessandro Innocenti and Angela Sirigu 6 Computational Intelligence Techniques for Trading and Investment Edited by Christian Dunis, Spiros Likothanassis, Andreas Karathanasopoulos, Georgios Sermpinis and Konstantinos Theofilatos Other books in the series include: Economics Lab An intensive course in experimental economics Alessandra Cassar and Dan Friedman Computational Intelligence Techniques for Trading and Investment Edited by Christian Dunis, Spiros Likothanassis, Andreas Karathanasopoulos, Georgios Sermpinis and Konstantinos Theofilatos First published 2014 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 711 Third Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2014 Selection and editorial matter, Christian Dunis, Spiros Likothanassis, Andreas Karathanasopoulos, Georgios Sermpinis and Konstantinos Theofilatos; individual chapters, the contributors The right of the editors to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data Computational intelligence techniques for trading and investment/edited by Christian Dunis, Spiros Likothanassis, Andreas Karathanasopoulos, Georgios Sermpinis and Konstantinos Theofilatos. pages cm Includes bibliographical references and index. 1. Investments–Mathematical models. 2. Investment analysis– Mathematical models. 3. Speculation–Mathematical models. 4. Computational intelligence. I. Dunis, Christian. HG4515.2.C66 2014 332.60285′63–dc23 2013036599 ISBN: 978-0-415-63680-3 (hbk) ISBN: 978-0-203-08498-4 (ebk) Typeset in Times New Roman by Wearset Ltd, Boldon, Tyne and Wear ‘To my family’ Christian Dunis ‘Dedicated to my daughter Kalliopi–Klelia’ Spiros Likothanassis ‘To my family’ Andreas Karathanasopoulos ‘To my beloved Fountouki’ Georgios Sermpinis ‘To my family and my beloved Zoi-Triantafyllia’ Konstantinos Theofilatos This page intentionally left blank Contents List of figures ix List of tables xi List of contributors xiii Preface xv Acknowledgements xix PART I Introduction 1 1 Computational intelligence: recent advances, perspectives and open problems 3 KONSTANTINOS THEOFILATOS, EFSTRATIOS GEORGOPOULOS, SPIROS LIKOTHANASSIS AND SEFERINA MAVROUDI 2 Financial forecasting and trading strategies: a survey 22 CHARALAMPOS STASINAKIS AND GEORGIOS SERMPINIS PART II Trading and investments with traditional computational intelligence techniques 41 3 Hidden Markov models: financial modelling and applications 43 SOVAN MITRA 4 Adaptive filtering on forecasting financial derivatives indices 66 CHRISTOS DIMITRAKOPOULOS, ANDREAS KARATHANASOPOULOS, GEORGIOS SERMPINIS AND SPIROS LIKOTHANASSIS viii Contents PART III Trading and investments with artificial neural networks 79 5 Modelling and trading the corn–ethanol crush spread with neural networks 81 CHRISTIAN L. DUNIS, JASON LAWS, PETER W. MIDDLETON AND ANDREAS KARATHANASOPOULOS 6 Trading decision support with historically consistent neural networks 116 HANS- JöRG VON METTENHEIM PART IV Trading and investments with hybrid evolutionary methodologies 131 7 Advanced short-t erm forecasting and trading deploying neural networks optimized with an adaptive evolutionary algorithm 133 KONSTANTINOS THEOFILATOS, THOMAS AMORGIANIOTIS, ANDREAS KARATHANASOPOULOS, GEORGIOS SERMPINIS, EFSTRATIOS GEORGOPOULOS AND SPIROS LIKOTHANASSIS 8 Using argumentation and hybrid evolutionary multi- model partitioning algorithms for efficient portfolio construction 146 NIKOLAOS SPANOUDAKIS, KONSTANTINA PENDARAKI AND GRIGORIOS BELIGIANNIS PART V Trading and investments with advanced computational intelligence modelling techniques 175 9 Forecasting DAX 30 using support vector machines and VDAX 177 RAFAEL ROSILLO, JAVIER GINER AND DAVID DE LA FUENTE 10 Ensemble learning of high- dimensional stock market data 192 MANOLIS MARAGOUDAKIS AND DIMITRIOS SERPANOS Index 216 Figures 1.1 The MLP neural network architecture 6 1.2 The RNN architecture 6 1.3 The HONN architecture 8 1.4 The RBF neural network architecture 9 1.5 Meta- heuristic algorithms flow chart 13 3.1 A diagram of a two-s tate Markov model with state transitions depicted by arrows 44 3.2 Graphs for the S&P 500 index 1976–1996 57 3.3 Graphs for the Nikkei 225 index 1976–1996 58 3.4 Graph of Bank of England interest rates (%) for each month from October 1980 to December 2008 59 3.5 Graph of interest rates states for each month from October 1980 to December 2008 60 5.1 The corn–ethanol crush CBOT daily closing prices 89 5.2 Histogram of corn–ethanol spread return series 90 5.3 A single output, inter-c onnected MLP model 96 5.4 Elman RNN architecture with two neurons/nodes for the hidden layer 97 5.5 Second order HONN with three inputs 99 6.1 Modelling with the HCNN. The initial hidden states can also be trained 118 8.1 The structure of the hybrid evolutionary multi- model partitioning algorithm used for financial time series forecasting 158 8.2 Mapping from a fixed dimensional internal representation to a variable length NARMAX parameter vector. The resulting order is n(p, q, r) = (3, 2, 3) 159 8.3 The fitness of each candidate model is the model conditional pdf 160 8.4 The PORTRAIT system architecture 163 8.5 The investment process using the PORTRAIT tool 164 8.6 A screenshot of the PORTRAIT tool 165 8.7 The bear market context, the moderate investor role and the specific context that combines them returns for all eight years 167