International Journal of Business and Finance Management Research
ISSN: 2053-1842
Vol. 8(3), pp. 33-47 April 2020
doi.org/10.33500/ijbfmr.2020.08.005



Application of artificial intelligence investment in capital markets: A case study of two constituent stocks of Dow Jones

Wei-Yuan Lin and Yuan-De Chu*

Department of Economics, Soochow University, Taipei, Taiwan.

*To whom correspondence should be addressed. E-mail: 0928chudavid@gmail.com.

Received 02 March, 2020; Received in revised form 13 April, 2020; Accepted 15 April, 2020.

Abstract


Keywords:
Reinforcement learning, Technical indices, Random forest, SVM, Prediction accuracy, Online investment strategy.

With the continuous advancement of science and technology, people continue to think about how to develop from weak artificial intelligence (AI) to strong AI. Researchers today already have big data richer than in the past, and use many new ideas of AI or data mining techniques to create many amazing achievements in various fields, such as the applications of sound, image and natural language. In addition, investment is also a hot topic that many people continue to explore, because investors always want to benefit from their decision-makings easily. In this paper, a reinforcement learning method is proposed to deal with the stocks market prediction. The information of the US stocks and data mining techniques were used to construct an online investment strategy, and conducted a comparative analysis of its different parameters. In this study, R software was apply to write the automated programs and build multiple technical indices or features. The shallow learning method--random forest (RF) was used to screen important features, and the support vector machine (SVM) classification method was also used to construct an online investment strategy. The empirical results obtained revealed that the average prediction accuracy can be increased significantly by adjusting the parameters of this model. Therefore, biology-based algorithms (GA, PSO and FOA etc.) or deep learning techniques (CNN, RNN etc.) can be further adopted in the future in order to obtain better accuracy and faster research results.

2020 BluePen Journals Ltd.







© 2013-2020 - BluePen Journals | Terms | Privacy policy | Advert | Help |