Research on Text Classification Based on TextRank

Extracting keywords from the result of word segmentation with the improved TextRank algorithm. Use the relative position of the words in the article to calculate the influence of position; the position of the coverage of the words and expressions is extended to the statement of the words and the key words as the feature of the text. Hadoop programming using naive Bayesian algorithm for text classification. The experiments show that the improved Textrank has a great improvement in classification performance, and the classification accuracy of naive Bayesian algorithm is 93% when the number of keywords is 40. Compared with the traditional, the accuracy rate increased by about 10%.