An improved random forest order review classification method based on Word2vec
Yuhan Wang, Xiaolan Xie, Bangjin Wang · 2022
In recent years, with the continuous development of the Internet, online shopping has gradually become a mainstream shopping method for people. At the same time, people often comment on the purchased goods, which expresses their shopping feelings on the one hand and provides corresponding reference for others' shopping on the other. In order to better mine and analyze the massive review data and extract valuable reviews from them, this paper proposes an improved random forest classification method based on Word2vec, which optimizes the number of decision trees and the number of leaf node variables in the random forest algorithm by incorporating the seagull algorithm to classify the crawled e-mall review data emotionally. The results show that the word2vec+ISOA-RF classification model is improved more significantly and can effectively enhance the performance of the classification algorithm.