Predicting Stock Price Using Random Forest Algorithm and Support Vector Machines Algorithm

This study develops a random forest and support vector machine model combined with various technical indicators for stock price analysis. This model uses these indicators as input features to analyze the data of Yuanda High Dividend (0056) in Taiwan, from December 26, 2007 to November 30, 2022. This research model can predict the next day stock price. In addition, the analysis of variance (ANOVA) is used as a feature screening method for the data, and the optuna package in python is used to find the optimal parameters for the two algorithms. We compare the accuracy of the two algorithms and verify the effectiveness of the feature screening method and parameter optimization. The research results show that the model after feature selection and parameter optimization, the results will have an average accuracy rate more than 55%. The prediction accuracy of the support vector machine (SVM) model using feature screening and parameter optimization is better than that of the random forest model.