Research on Micro-blog Text Presentation Model Based on Word2vec and TF-IDF
Qiu Yan, Bo Yang · 2021 IEEE Asia-Pacific Conference on Image Processing, Electronics and Computers (IPEC) · 2021
It is of great practical significance to mine valuable information from micro-blog data, assist government departments to manage network environment and find potential business opportunities for enterprises. The clustering accuracy of topic detection depends on the quality of the text representation model. Micro-blog is short, scattered, data redundancy and so on. Therefore, traditional text representation technology methods will lead to high dimensional sparsity of semantic matrix. And the semantic information related to the topic is lost, resulting in insufficient clustering effect. In this paper, a text representation model based on the improved Word2vec and TF-IDF is constructed, and it is applied to topic detection. The comparative experiment shows that the Word2vec &TF-IDF model has a better effect of text expression and improves the accuracy of micro-blog text clustering.