Modeling and analysis of financial markets using system adaptation and frequency domain approach

In this paper, we adopt the concept of feedback systems to model the behavior of financial markets, or more specifically, the stock market. Based on a feedback adaptation scheme, we model the price movement of a stock counter with an internal model using the well-known exponential moving average technique together with an adaptive filter, which is to be adjusted for an individual counter. Its input-output behavior, and internal as well as external forces are then identified. The estimated external force is analyzed and characterized in the frequency domain. Test results show that there are highly significant components contained in its frequency contents. The appearing time and locations of these components provide empirical indications of majoring turning points in the market and seasonality effects in stock returns. Statistical tests such as Bai-Perron test, Kruskal-Wallis H test and Mann-Whitney U test are employed to evidence these market properties.