Research On Topic Recommendation Of Weibo Based On Knowledge Graph Construction

Due to the rapid increase in the number of Weibo users,the number of topics generated on the Weibo platform every day is also large.In the face of a large number of generated topics,it is often difficult for users to screen out the topics they are interested in.To solve this problem,this paper proposes a Weibo topic recommendation algorithm based on knowledge graph to help users find the interesting topics.The main research contents are as follows:(1)Establish Weibo topic knowledge graph.First,the acquired Weibo text is preprocessed,and the processed Weibo text is obtained by using the learned suffix rules to obtain the named entity.Then the dependency syntax analysis of Weibo statements is carried out,and the Weibo relationship is extracted by Bootstrapping Weibo relation extraction algorithm.Finally,the named entity is regarded as the node,and the extracted corresponding relationship is regarded as the edge connecting the two nodes,and the visual display of the knowledge graph is realized by software drawing.(2)Set up the topic user interest degree matrix.Define the user Weibo word feature set,and use TF-IDF to obtain it according to the user Weibo history data.The Weibo topic knowledge graph is matched with the Weibo topic knowledge graph,and the user interest matrix is obtained,and the preference of all users for all topics is obtained.At last,the user is analyzed by the method of partition clustering(k-means).(3)Weibo recommendation based on topic knowledge spectrum and user interest.We first define the Weibo topic naming entity coefficient,which is used to characterize the importance of Weibo topic named entity to users,and then on the basis of Weibo topic knowledge graph and user cluster analysis,The Weibo topic is recommended based on the topic knowledge graph,and the Weibo topic recommendation set is obtained.Then the similarity between the weight vector of Weibo topic feature words to be recommended and the named entity coefficient of Weibo topic is calculated.Based on the screening of user content,the recommended topic set with high similarity is obtained,and finally the recommendation set is obtained..Through the verification of the experimental data,the Weibo recommendation algorithm based on the topic knowledge graph can improve the accuracy of the topic recommended to the user to a great extent,and greatly reduce the time for the user to find the topic of interest.Timely and efficient help users to obtain useful information for themselves.This paper innovatively combines Weibo topic knowledge graph with collaborative filtering recommendation,alleviates the cold start problem of collaborative recommendation to a great extent by establishing user interest matrix,and defines Weibo topic naming entity coefficient.The topics obtained by collaborative filtering are filtered out of the topics that do not conform to the user's preferences,and the accuracy of the recommendation is finally improved.Figure[13]table[16]reference[62].