Machine Learning Model-Based Financial Market Sentiment Prediction and Application

Abstract

Academic interest in stock market forecasting has risen dramatically in recent years. An in-depth understanding of stock movement data and strong analytical abilities are prerequisites for making reliable stock market forecasts. This responsibility calls for innovative methods of execution to provide the highest level of accuracy in the prediction and, by extension, the best potential profits for investors. The major emphasis of this article is on developing a robust stock market forecasting system by using effective machine learning techniques. The research has three parts: (1) stock market data set preprocessing; (2) application of two supervised machine learning algorithms, K-Nearest Neighbor (KNN) and Random Forest (RF); and (3) evaluation of the accuracy and efficacy of predicting for the two recommended models. The experimental findings showed, while both recommended models performed well in terms of accuracy ratio (with values of 90.23 percent, 91.12 percent, and 90.17 percent for precision, recall, and F-measure, respectively).