An intelligent system for discovering mathematical models for financial time series prediction

Abstract

We describe a computer program that can be considered an intelligent system for the domain of financial time series prediction. Given a financial time series for a specific problem, the intelligent system develops mathematical models for the problem based on the geometry of the data, using different statistical methods. The computer program develops the kind of models that are more likely to give a good prediction, based on the knowledge of the human experts contained in the knowledge base. The intelligent system then analyzes the mathematical models obtained before to make a selection of the model that will give us the best prediction. This selection is done by the intelligent system using a combination of heuristics and calculations that are contained also in the knowledge base. An intelligent system that can learn models from financial data would be very useful in practice in making the job of prediction more easy and less time consuming. >