How Efficient is Machine Learning on Stock Exchange Expectations?

In this study, it is aimed to compare the performances of the methods by predicting the movement directions of stock market indexes in developed countries by employing machine learning methods (MLMs) and determining the best estimation method. For this purpose, the movement directions of indexes such as NYSE 100 (the USA), NIKKEI 225 (Japan), FTSE 100 (the UK), CAC 40 (France), DAX 30 (Germany), FTSE MIB (Italy) and TSX (Canada) were estimated by employing the decision tree, random forest k-nearest neighbor, naive Bayes, logistic regression, support vector machines, and artificial neural network methods. According to the obtained results, the best method for estimating the movement directions of stock market indexes was determined as the artificial neural networks. Moreover, the artificial neural networks, which is the best method, were determined invalid for all indexes, in other words, it was not the method with the highest accuracy performance in all indexes