Biologically Inspired Algorithms for Financial Modelling

by Brabazon, Anthony; O'Neill, Michael

• ISBN13: 9783540262527
• ISBN10: 3540262520
• Format: Hardcover
• Copyright: 2006-03-30
• Publisher: Springer-Verlag New York Inc

Summary

Predicting the future for financial gain is a difficult, sometimes profitable activity. The focus of this book is the application of biologically inspired algorithms (BIAs) to financial modelling.In a detailed introduction, the authors explain computer trading on financial markets and the difficulties faced in financial market modelling. Then Part I provides a thorough guide to the various bioinspired methodologies - neural networks, evolutionary computing (particularly genetic algorithms and grammatical evolution), particle swarm and ant colony optimization, and immune systems. Part II brings the reader through the development of market trading systems. Finally, Part III examines real-world case studies where BIA methodologies are employed to construct trading systems in equity and foreign exchange markets, and for the prediction of corporate bond ratings and corporate failures.The book was written for those in the finance community who want to apply BIAs in financial modelling, and for computer scientists who want an introduction to this growing application domain.

Table of Contents

Introduction	1	(14)
Biologically Inspired Algorithms	2	(2)
Artificial Neural Networks	2	(1)
Evolutionary Computation	2	(2)
Social Systems	4	(1)
Artificial Immune Systems	4	(1)
Computer Trading on Financial Markets	4	(1)
Challenges in the Modelling of Financial Markets	5	(3)
Do Prices Follow a Random Walk?	6	(1)
Attack of the Anomalies	7	(1)
Linear Models	8	(2)
Structure of the Book	10	(5)
Part I Methodologies
Neural Network Methodologies	15	(22)
A Taxonomy of NNs	15	(1)
The Multi Layer Perceptron	16	(13)
Training an MLP	20	(3)
Practical Issues in Training MLPs	23	(5)
Recurrent Networks	28	(1)
Radial Basis Function Networks	29	(3)
Self-organising Maps	32	(3)
Implementing a SOM	33	(2)
Summary	35	(2)
Evolutionary Methodologies	37	(36)
Genetic Algorithm	37	(12)
Canonical GA	40	(1)
Example of the GA	41	(2)
Extending the Canonical GA	43	(5)
Schema and Building Blocks	48	(1)
Differential Evolution	49	(5)
DE Algorithm	49	(5)
Genetic Programming	54	(9)
More Complex GP Architectures	58	(5)
Combining EA and MLP Methodologies	63	(5)
Applying EAs to Evolve Trading Rules	68	(2)
Recent Developments in Evolutionary Computation	70	(1)
Summary	71	(2)
Grammatical Evolution	73	(16)
Grammatical Evolution	73	(9)
Biological Analogy	74	(2)
Mapping Process	76	(3)
Mapping Example	79	(3)
Mutation and Crossover in GE	82	(2)
Recent Developments in GE	84	(4)
Search Engine	84	(1)
Meta-grammars	85	(2)
πGE	87	(1)
Applications and Alternative Grammars	87	(1)
Summary	88	(1)
The Particle Swarm Model	89	(10)
PSO Algorithm	89	(5)
Constriction Coefficient Version of PSO	92	(1)
Parameter Settings for PSO	93	(1)
Discrete PSO	94	(1)
Comparing PSO and the GA	94	(1)
MLP-Swarm Hybrids	95	(1)
Grammatical Swarm	95	(1)
Example of a Financial Application of PSO	96	(1)
Recent Developments in PSO	96	(1)
Summary	97	(2)
Ant Colony Models	99	(8)
Ant-Foraging Models	99	(5)
Ant-Foraging Algorithm	100	(4)
A Financial Application of ACO	104	(1)
Ant-Inspired Classification Algorithms	105	(1)
Hybrid Ant Models	105	(1)
Summary	106	(1)
Artificial Immune Systems	107	(14)
Overview of Natural Immune Systems	108	(5)
Innate vs Adaptive Immunity	108	(1)
Components of the Immune System	108	(5)
Designing Artificial Immune Algorithms	113	(3)
Negative Selection Algorithm	113	(1)
Clonal Expansion and Selection Algorithm	114	(2)
Financial Application of the Negative Selection Algorithm	116	(2)
Summary	118	(3)
Part II Model Development
Model Development Process	121	(22)
Project Goals	121	(3)
What to Forecast?	121	(2)
What Performance Measure Is Appropriate?	123	(1)
Data Collection	124	(6)
Trading Philosophy	124	(4)
How Much Data Is Enough?	128	(2)
Selecting and Preprocessing the Data	130	(4)
Selection	130	(1)
Preprocessing	130	(4)
Postprocessing the Output	134	(1)
Entry Strategy	134	(1)
Exit Strategy	134	(1)
Money Management	135	(1)
Validating the System	135	(5)
Implementation and Maintenance	140	(2)
Summary	142	(1)
Technical Analysis	143	(16)
Technical Indicators	144	(10)
Moving Average	146	(2)
Momentum	148	(1)
Breakout	149	(1)
Stochastic Oscillators	150	(2)
Volume Data	152	(1)
Other Indicators	153	(1)
Using Technical Indicators in a Trading System	154	(1)
Summary	155	(4)
Part III Case Studies
Overview of Case Studies	159	(2)
Index Prediction Using MLPs	161	(14)
Methodology	162	(7)
Model Selection	166	(1)
Model Stacking	167	(2)
Results	169	(3)
RMSE and Correlation	169	(2)
Trading System	171	(1)
Discussion	172	(3)
Index Prediction Using a MLP-GA Hybrid	175	(8)
Methodology	175	(3)
Model Construction	176	(2)
Results	178	(4)
MLP-GA	179	(1)
Analysis of Weight Vectors	180	(2)
Discussion	182	(1)
Index Trading Using Grammatical Evolution	183	(10)
Methodology	183	(6)
GE System Setup	188	(1)
Results	189	(1)
Discussion	190	(3)
Adaptive Trading Using Grammatical Evolution	193	(10)
Introduction	193	(1)
Methodology	193	(3)
Moving Window	194	(1)
Variable Position Trading	194	(1)
Return Calculation	195	(1)
Results	196	(5)
Training Returns	197	(2)
Out-of-Sample Returns	199	(2)
Discussion	201	(2)
Intra-day Trading Using Grammatical Evolution	203	(8)
Background	203	(1)
Methodology	204	(4)
Trading System	206	(1)
GE System Setup	207	(1)
Results	208	(2)
Discussion	210	(1)
Automatic Generation of Foreign Exchange Trading Rules	211	(8)
Background	211	(1)
Methodology	212	(2)
Results	214	(4)
US-STG	216	(1)
US-Yen	217	(1)
US-DM	217	(1)
Discussion	218	(1)
Corporate Failure Prediction Using Grammatical Evolution	219	(10)
Background	220	(2)
Definition of Corporate Failure	220	(1)
Explanatory Variables	221	(1)
Methodology	222	(2)
GE System Setup	223	(1)
LDA Method	224	(1)
Results	224	(2)
Form of the Evolved Classifiers	225	(1)
Discussion	226	(3)
Corporate Failure Prediction Using an Ant Model	229	(10)
Background	229	(1)
Methodology	230	(5)
Ant System	231	(4)
Results	235	(3)
Discussion	238	(1)
Bond Rating Using Grammatical Evolution	239	(10)
Background	240	(1)
Rating Process	240	(1)
Methodology	241	(2)
Results	243	(4)
Discussion	247	(2)
Bond Rating Using AIS	249	(6)
Methodology	249	(3)
Results	252	(1)
Discussion	252	(3)
Wrap-up	255	(2)
References	257	(14)
Index	271