Research on stock prediction method based on random forest

The complexity and nonlinearity of stock prices have made the study of stock price prediction techniques a significant area of concern in the finance industry. This paper approaches the stock price prediction issue as a categorization challenge, considering the stock's opening, minimum, maximum, trading volume, and daily fluctuation amplitude as key factors. The stock's closing price on the same day is forecasted using logistic regression, support vector machine, decision tree, and random forest algorithms, followed by a comparative analysis of the predictive outcomes from these models. The lattice search is then used to optimize the random forest model. According to the experimental results, the random forest model outperforms the other methods in terms of stock price prediction ability. Following grid search optimization, the model's stability and prediction accuracy are further enhanced.