Design And Implementation Of Microblog News Position Detection System

With the rapid development of the Internet and the widespread application of many social platforms,the content of social media platforms is gradually showing a trend of diversification.Weibo is a social media platform frequently used by users,on which users can initiate topics to attract a large number of people to participate in the discussion.There will be a lot of text content on the microblog platform at any time,most of which are tagged with topic tags,and these microblog books usually contain valuable information about a target topic(such as the direction of public opinion after hot news events).Therefore,in the face of a large amount of microblog data,It is particularly necessary to explore relevant research on the potential position of users.At present,when people conduct research on stance detection,the key point of research usually focuses on the direction of analyzing the text of micro-blog,and little attention is paid to the influence of topic information on stance detection.Subsequently,some improvements and adjustments were made to address the shortcomings of the research.The main improvement method was to combine the topic information with the microblog text.The method mainly consists of simple splicing of two text sequences,or using attention-based calculation to decentralize the topic information into the microblog information.In this paper,a position detection algorithm model based on Condition-CNN-BILSTM is proposed,which not only combines the text content and topic information,but also expands the content of topic information by extracting phrases in the text,and significantly improves the weight of topic information in position detection task:The main work of this paper is as follows:1)Research position detection algorithm based on deep learning.In the face of a large amount of text information in microblog news,this paper proposes a microblog news position detection algorithm based on Condition-CNN-BiLSTM.Firstly,the corresponding phrases in the microblog news text are extracted to form the topic set.Then,Bert is applied to obtain the vector representation of the topic set and microblog news text.In addition,this paper proposes the construction of Condition matrix between two sentence vectors,and applies the deep learning model of CNN-BiLSTM fusion to the Condition layer to complete the extraction of the relationship features between microblog news and topic sets,so as to obtain the final position information and complete the purpose of microblog news position detection.Then the experimental comparison with other excellent weibo position detection models on the data set proves the actual effectiveness of the proposed algorithm in position detection task.2)Completed the design and implementation of microblog news position detection system.Based on the deep learning model of CNN-BiLSTM-Condition,a microblog news position detection system is designed.To ensure that users input the topic to be detected and the start and end time of the query in the system,the background program can climb all the comment text of the topic within the specified period and analyze it by applying the microblog news position detection model designed in this paper.Finally,the position detection result of the topic can be displayed on the system page.In addition,after the system is built,test cases are designed to test the system’s functions and parallelism ability to ensure the safe and stable operation of the system.