Stock Market Modeling: A System Adaptation Approach

Abstract

This thesis aims to depict and analyze the financial markets using a general framework based on systems theory. A system adaptation framework has been proposed for modeling the stock market, or more generally, the financial market from a dynamic system point of view. The movement of stock market indices is modeled within this framework that is composed of an internal dynamic model and a time-varying adaptive filter. As the inputs are essential, a double selection method based on both empirical and statistical knowledge is proposed to select the influential factors of a specific market. Through its design and influential factors, this framework provided excellent prediction results and more meaningful insights into the behavior of the stock market. An application of this framework has also been introduced which focused on the forecasting of major market turning periods. It provided an accurate forecast by using the frequency pattern-based identification enhanced by the detection of the instability in the internal model. To facilitate the analysis of the stock market, a MATLAB toolkit has been developed.