Stock Selection with a Novel Sigmoid-Based Mixed Discrete-Continuous Differential Evolution Algorithm

Author(s): M. George Fernandez, V.Angalin priyadharshini

Abstract: Stock market is a mostly a virtual exchange of securities and Risk that is, shares for raising the finance and derivatives of a leading companies. Predicting the stock trend is highly challenging since the stock market data are highly time variant data and nonlinear pattern. To introspect challenges in stock market need to overcome the impediments and strive for further improving our focus on prediction of share market. In this paper investigated a stock selection model by introducing and improving the Differential Evolution (DE) algorithm for model optimization. Two main steps are involved in this stock selection models. The selection models are stock scoring and stock ranking. First, a stock scoring mechanism is designed, in which stocks were evaluated based on various fundamental and technical features. Second, the top-ranked stocks are selected to formulate an equal-weighted portfolio as the model output. For choosing proper features and optimizing the corresponding weights. This investigation of stock selection model with a novel sigmoid-based DE algorithm for the mixed discrete-continuous optimization, and to verify its superiority over benchmark models with other model designs (in terms of different decision variables and fitness functions) and other popular optimization techniques.