Multinational Effects of Foreign Exchange Rate in Stock Index with Classification Models for Medium-term Investment

This research builds prediction models based on classification algorithms to propose a novel method, which provides research and practical guidelines and answers research questions we proposed in this field. Based on extant classification approaches and their limitations, the proposed method integrates multinational stock index and foreign exchange rate of main trading countries and builds effectively self-learning models to adjust behaviors of the medium-term investment dynamically. The proposed approach is unique in several aspects. First, the classification algorithms approach, a type of machine learning technologies, automatically generates patterns of medium-term stock index trend based on big data analysis. The method overcomes the problem of medium-term investment risks. Second, we evaluate foreign exchange rate to prove that it is a significant factor for stock index. Third, incorporating foreign exchange rate into multinational stock index has significant improvement on accuracy of prediction. This paper utilizes popular machine learning algorithms such as SVMs to improve the effectiveness of the proposed method. The results of the evaluation via a medium -term data analysis indicate that the approach shows advantages in the accuracy of stock index prediction in comparison with existing methods only considering stock index. JEL classification numbers: G10