Biologically Inspired Algorithms for Financial Modelling
暂无分享,去创建一个
Anthony Brabazon and Michael O’Neill of the University College Dublin have just published an interesting book that introduces a wide range of biologically inspired algorithms and their applications in financial modelling. It is divided into three parts, containing in total 21 chapters. The largest, Part I, contains brief introductory explanations of three forms of artificial neural networks (multi-layer perceptrons-MLP, radial basis function network, and self-organising maps), evolutionary computation algorithms (GA, differential evolution, genetic programming, and combinations with MLP), social systems (particle swarm optimization, ant colony models, and their MLP hybrids) and artificial immune systems. The smallest of the three is Part II. It focuses on the development of market trading systems. Topics addressed range from goal determination and data collection to model construction and validation. A short chapter is also devoted to technical analysis of equity markets. In Part III, ten case studies are reported. Four of them investigate stock market index predictions, two cases focus on stock returns predictions and trading, and the remaining four tackle prediction of exchange rates, corporate failure and bond ratings. This book is a well-written, easy to read, brief introduction to state-of-the-art biologically inspired algorithms. It documents their potential as ‘modeling for prediction and classification techniques’ in financial markets. The reader will find most chapters rather succinct. There are no more than a few simple mathematical equations to follow in most of them. The few lines of programming code are mainly in Chapter 4 on Grammatical Evolution. Readers who have had no exposure to any of the algorithms or have no background in finance and financial trading strategies, will find some of the material rather challenging if not intimidating. Readers without prior knowledge in both are advised to read this book after completing a more focused training on one of the algorithms and an exposure to financial trading and jargon. Without these two backgrounds, the book will be difficult to follow. The authors’ statement “No prior knowledge of either biologically inspired algorithms or financial modelling is assumed” is unrealistic. Readers with programming skills will find the first half of the book easy to follow and Parts II and III more difficult to understand without completing additional reading, e.g. from the book’s references. Readers with financial trading strategies background will find Parts II and III too simple (and perhaps rather weak) but they may be interested in the algorithms in Part I.
[1] P. Kaufman. New Trading Systems and Methods , 2005 .
[2] Abdol S. Soofi,et al. Modelling and Forecasting Financial Data , 2002 .
[3] Lakhmi Jain,et al. Computational Economics: A Perspective from Computational Intelligence , 2006 .
[4] P. McNelis. Neural networks in finance : gaining predictive edge in the market , 2005 .