Research and Implementation of SVM and Bootstrapping Fusion Algorithm in Emotion Analysis of Stock Review Texts

Abstract

In view of the unclear tendency of text sentiment in stock review texts, this paper constructs a stock evaluation text sentiment analysis algorithm that combines SVM and Bootstrapping. Firstly, to obtain the data series of the stock review text to be processed, a web crawler is used to collect the content of the financial blogger's stock review website in the first half of 2019. Then the SVM algorithm is used to classify the stock review text to obtain the emotional feature words. The set is then reconstructed using the Bootstrapping algorithm to obtain a high-performance classifier. Finally, the model of the analyzed emotional feature words is evaluated. The experimental results show that compared with the traditional algorithm, the recall rate is increased by 3.3%, the accuracy rate is improved by 3.9%, and the weighted harmonic average is improved by 3.9%. The improved algorithm classification is better than the traditional one. The accuracy and recall rate are better, and the weighted harmonic mean has been greatly improved.