An Evolutionary Bootstrap Method for Selecting Dynamic Trading Strategies

This paper combines techniques drawn from the literature on evolutionary optimization algorithms along with bootstrap based statistical tests. Bootstrapping and cross validation are used as a general framework for estimating objectives out of sample by redrawing subsets from a training sample. Evolution is used to search the large space of potential network architectures. The combination of these two methods creates a network estimation and selection procedure which aims to find parsimonious network structures which generalize well. Examples are given from financial data showing how this compares to more traditional model selection methods. The bootstrap methodology also allows more general objective functions than usual least squares since it can estimate the in sample bias for any function. Some of these will be compared with traditional least squares based estimates in dynamic trading settings with foreign exchange series.

[1]  Robert Tibshirani,et al.  A Comparison of Some Error Estimates for Neural Network Models , 1996, Neural Computation.

[2]  K. Lewis,et al.  Puzzles in International Financial Markets , 1994 .

[3]  Robert Tibshirani,et al.  An Introduction to the Bootstrap , 1994 .

[4]  B. LeBaron.  Forecast improvements using a volatility index , 1992 .

[5]  Xin Yao,et al.  A review of evolutionary artificial neural networks , 1993, Int. J. Intell. Syst..

[6]  C. Engel,et al.  The Forward Discount Anomaly and the Risk Premium: A Survey of Recent Evidence , 1995 .

[7]  D. Fogel.  Evolutionary algorithms in theory and practice , 1997, Complex..

[8]  J. Shao.  Linear Model Selection by Cross-validation , 1993 .

[9]  Dorothea Heiss-Czedik,et al.  An Introduction to Genetic Algorithms. , 1997, Artificial Life.

[10]  Martin Eichenbaum,et al.  Some Empirical Evidence on the Effects of Shocks to Monetary Policy on Exchange Rates , 1995 .

[11]  Peter Hall,et al.  Using the bootstrap to estimate mean squared error and select smoothing parameter in nonparametric problems , 1990 .

[12]  Gerhard Paass,et al.  Assessing and Improving Neural Network Predictions by the Bootstrap Algorithm , 1992, NIPS.

[13]  Halbert White,et al.  Artificial neural networks: an econometric perspective ∗ , 1994 .

[14]  B. Efron.  Estimating the Error Rate of a Prediction Rule: Improvement on Cross-Validation , 1983 .

[15]  Robert Tibshirani,et al.  Model Search and Inference By Bootstrap "bumping , 1995 .

[16]  H. White,et al.  A Reality Check for Data Snooping , 2000 .

[17]  Vasant Honavar,et al.  Evolutionary Design of Neural Architectures , 1995 .

[18]  R. Sweeney,et al.  Beating the Foreign Exchange Market , 1986 .

[19]  G. Schwarz.  Estimating the Dimension of a Model , 1978 .

[20]  Jasmina Arifovic.  Genetic algorithm learning and the cobweb model , 1994 .

[21]  Blake LeBaron,et al.  A Bootstrap Evaluation of the Effect of Data Splitting on Financial Time Series , 1996, IEEE Trans. Neural Networks.

[22]  Richard M. Levich,et al.  The Significance of Technical Trading-Rule Profits in the Foreign Exchange Market: a Bootstrap Approach , 1991 .

[23]  Andreas S. Weigend,et al.  Nonlinear Trading Models Through Sharpe Ratio Maximization , 1997, Int. J. Neural Syst..

[24]  W. Press,et al.  Numerical Recipes: The Art of Scientific Computing , 1987 .

[25]  Clive W. J. Granger,et al.  A Decision_Theoretic Approach to Forecast Evaluation , 2000 .

[26]  B. LeBaron.  Technical Trading Rules and Regime Shifts in Foreign Exchange , 1991 .

[27]  D. Bigman,et al.  Exchange rate and trade instability: Causes, consequences, and remedies , 1986 .

[28]  Lizhong Wu,et al.  Optimization of trading systems and portfolios , 1997, Proceedings of the IEEE/IAFE 1997 Computational Intelligence for Financial Engineering (CIFEr).

[29]  Lee R. Thomas.  The Currency-hedging debate , 1990 .

[30]  Stephen L Taylor,et al.  Rewards Available to Currency Futures Speculators: Compensation for Risk or Evidence of Inefficient Pricing? , 1992 .