The Nikkei Stock Average Prediction by SVM

Abstract

The problem of how to extract structures hidden in large amounts of data is called “data mining”. Using a support vector machine (SVM), which is one of the data mining methods, I predicted the rise and fall of the Nikkei stock average one day, one week, and one month later. As explanatory variables, we used the historical rate of change in US stock prices and the Nikkei Stock Average. As a result of the analysis, it was possible to stably improve the prediction accuracy of the diary average stock price one day later compared to random prediction. In addition, SHAP was used to analyze whether the explanatory variables were appropriate. As a result, we found that the effect of each explanatory variable on the analysis results differs depending on how the training set and test set are divided. We made it a future task to make stock price predictions using SVMs more concrete and convincing.