Research on intelligent classification of work orders based on TFIDF_SP algorithm and LDA-Word2vec model

In order to improve the accuracy of intelligent classification of 95598 tickets, this paper proposes an intelligent classification method based on TFIDF_SP algorithm and bLDA topic. In view of the shortcomings of the TFIDF algorithm, an improved TFIDF_SP algorithm was designed to distinguish the importance of feature words in different parts of the document, and the effectiveness of the algorithm was judged according to the comparison results of different feature selection methods. The TFIDF_SP algorithm and the traditional TFIDF algorithm were used to extract the feature words in the database, and the K_means algorithm was used for clustering, and the comparison results showed that the clustering effect of the feature words extracted by the TFIDF_SP algorithm was better than that of TFIDF, which verified the reliability of the proposed method. Simulation experiments are carried out on 95598 work orders, and the proposed TFIDF_SP algorithm and LDA-Word2vec model improve the clustering accuracy by 1.81% compared with the traditional method, which verifies that the TFIDF_SP and LDA-Word2vec models are more suitable for the classification research of handling work order complaints.