Stock market trend prediction using dynamical Bayesian factor graph

Abstract

A new qualitative method using the concept of dynamical Bayesian factor graph is developed in this paper for the prediction of stock market trend. The essence of this method is to compute the corresponding dynamical Bayesian factor graph for a selected set of macroeconomic factors over a period of time of interest. The computed time series of graphs capture both the mutual influential relationships and the evaluation of these relationships among these factors over a specified period of time. Then any topological structural change in the adjacent graphs at anytime predicts a change in market trend in a short future. Our computational analysis also indicates that if the topological structure of the underlying dynamical Bayesian factor graph is unchanged, the general market trend appears invariant. We demonstrate the effectiveness of this method by applying it in the analysis of market trends in the US stock market and Chinese stock market with good prediction results.