Stock Selection by Using an Improved Quick Artificial Bee Colony Algorithm

Abstract

Computation Intelligence have inspired many researchers to develop the capability of computer to learn and solve the complex task in real-world problems. In this work, we proposed a Artificial Bee Colony (ABC) to deal with the Stock Selection problem. We apply a Sigmoid-based Discrete-Continuous with ABC to select appropriate features for stock scoring. The empirical study tests the performance of ABC compare with Genetic Algorithm (GA) and Differential Evolution (DE) algorithm by using data from the Stock Exchange Thailand. The empirical results show that the novel model stock selection significantly outperforms in terms of both investment return, diversity and model robustness.