Research On Structural Data Recognition Technology Of News Text And Its Application In Key Information Extraction Of Quality Supervision News

Shanghai Institute of quality supervision and inspection technology needs to obtain the key information of commodity name,attribute,qualified batch,risk factors and so on in the quality inspection results of consumer goods in various provinces and cities of China.These information are published on the quality supervision news websites of various regions.Each news has a detailed description of a quality supervision inspection action,and the narrative recording form makes each news more lengthy Long,covering such as quality inspection personnel introduction,commodity prices,illegal handling opinions and other needs of information,and the above news text does not have a unified writing standard,Shanghai Institute of quality supervision and inspection technology needs to send people to read and count the new news notice one by one every week,there is a lot of tedious duplication of labor,in view of the above situation,this paper designs and develops A key information extraction system is designed to solve this problem.In the development of key information extraction system to solve the problem of key information extraction in quality supervision news,this paper completes the following work(1)Through the web crawler crawling the news text data of the quality supervision news network all over the country to extract the key information.(2)In the data preprocessing stage,the key sentence extraction technology is used to focus on the key information and reduce the corpus.By comparing the performance of key sentence extraction technology based on Text Rank algorithm and deep learning model in quality supervision news text,the output of Text Rank algorithm is selected as the result of data preprocessing.(3)In the selection of key information extraction algorithm,through the experimental comparison based on TF-IDF,Text Rank,and the combination of knowledge mapping and deep learning,we finally choose the combination of knowledge mapping and deep learning as the core algorithm of the follow-up key information extraction system.(4)Through the system analysis and architecture design,using the front-end technologies such as Vue,layui,bootstrap and the back-end technologies such as C #,python,the key information extraction system of quality supervision news is developed,which realizes the functions of automatic crawling,key sentence extraction,key information extraction,data query and visualization of quality supervision news,and solves the problem of Shanghai quality supervision through the above system The demand of information extraction in Laboratory Technology Research Institute.The research results and the developed system have been put into use in Shanghai Institute of quality supervision and inspection technology.Through the key information extraction system,the shortage of manual extraction has been well solved,and the work efficiency has been improved.