Research On Topic Extraction Algorithm Of Social Insurance And Housing Fund

With the rapid development of China's economy,its attention to people's livelihood is constantly improving.As an essential part of it,the social insurance and housing fund policy is gradually improving.Therefore,there will be many policy adjustments in the country and local governments,and each policy adjustment corresponds to the issuance of policies and regulations.Suppose we can extract and analyze the information contained in the policies and regulations in the field of social insurance and housing fund through relevant technologies.In that case,we can provide strong data support for the research in the field of social insurance and housing fund.Finding accurate and practical information in a large number of texts has become one of the hot research issues in the fields of natural language processing,machine learning,and artificial intelligence.The topic is an important information carrier of the text,so as long as we accurately find the topic of the text,we can save time and effort to obtain the main content of the text.This paper proposes a dual label topic extraction algorithm BL-LDA based on Latent Dirichlet Allocation in the field of social insurance and housing fund.The algorithm proposed in this paper adds audit words information and regional economic information to the Latent Dirichlet Allocation.The algorithm uses graph analysis technology to weigh the words in the audit method to find its keywords.It compares the keywords with the words in the social insurance and housing fund policies and regulations to mark whether the words are audit words.At the same time,it uses per capita G.D.P.as the standard to classify the regions.The words of the text are marked as developed or underdeveloped,and the two kinds of information are added to LDA to extract the topic of the text.Finally,the experimental results show that BL-LDA achieves better performance,which shows that the improved topic extraction algorithm can better adapt to the social insurance and housing fund policy and regulation texts.After clustering the extracted topics and using TFIDF to get the keywords of the topics,the application value of BL-LDA is demonstrated by analyzing the keywords.