Research on the Improved Word2Vec Optimization Strategy Based on Statistical Language Model

In order to improve the text matching degree and calculation accuracy of the short text classification method, this paper studies the optimization of the short text classification method of the Word2Vec algorithm. First of all, in order to solve the problem of insufficient features extracted by traditional text sign extraction methods, this paper constructs the feature vector space of the text through the word vector model of distributed semantic expression. Secondly, it uses the Word2Vec algorithm to train the word vector database, and extracts the feature vectors in this paper through the statistical language model. At last, it verifies the reliability and practicability of the algorithm proposed in this paper through empirical analysis. The research results show that the word vector semantic model based on distributed semantic expression can establish the feature vector space of short texts, reducing the probability of gradient explosion. The use of statistical language models can calculate the feature vector of each word combination in the short text. Thus, the feature vector value based on the word sequence is obtained. The Word2Vec algorithm improved by the empirical analysis model has a good classification effect on short text, which proves the effectiveness of the method of this research. This study proved that the Word2Vec algorithm can be optimized by using the statistical language model to improve the text matching degree and calculation accuracy of the text classification algorithm.