A Novel Text Classification Approach Based on Word2vec and TextRank Keyword Extraction

In the era of big data, a practical challenge is how to extract the keywords and classify the content of the massive text data efficiently. To address this challenge, through extensive research of existing approaches, we propose a novel text classification approach based on TextRank and Word2vec. In our approach, we integrate the external influence of the Word2vec training model and the location-aware weighting influence of the keywords into the Textrank module. Furthermore, we propose a novel text re-classification algorithm based on the combination of the keywords and KNN. The experiment results show that our approach achieves better accuracy and performance compared to TF-IDF and TextRank algorithm.