Linear and nonlinear structures in predicting stock return

Predicting stock return is possible with discovery generating process of stock price behavior patterns. The success rate in the discovery of these patterns، indicate the efficacy of the prediction. In other words، stock price generating process can be studied as a dynamic model. This process may be obtained in linear model forms، nonlinear model forms or stochastic model forms. This study describes linear predictor structures in the form of capital asset pricing model and Fama and French three factor model and nonlinear structures in the form of artificial neural networks.