Design And Implementation Of Chinese Journal Data Analysis System Based On Text Mining

With the development of science and technology goes rapidly,researchers' number in various fields continues to grow,more and more scientific and technological achievements have been created,the results for the researchers to bring new methods and new knowledge,but new problems follows.For the academic journal library of CNKI,there are more than 8820 Chinese academic journals,among which more than 1960 are the core journals of Peking University,and the daily update volume of papers is more than 5000.It can be say that,the journal platform has become a very important way for researchers to publish research achievements and access literature.If researchers are unable to conduct complete and rapid data analysis on massive journal resources,it will reduce the researchers' research and utilization of literature,and so,in order to realize the research and analysis of such huge journal paper data,this paper proposes a text mining-based Chinese journal data analysis system for Chinese journal papers.The functional modules of the Chinese journal data analysis system are divided into:statistical analysis,text similarity analysis,hot spot analysis and hot spot prediction,the research content is CNKI's 2010?2019 Chinese core journal papers' titles,authors,keywords,abstracts,funding,publication time,location and unit.The K-Means clustering algorithm,data mining Apriori algorithm,time-series multi-word keyword co-occurrence prediction algorithm and the topic extraction model LDA algorithm provide researchers with statistical analysis services,similarity analysis service,hot spot analysis service and hot spot prediction service.The Chinese journal data analysis system adopts Python language and Django framework as a whole,and uses crawler technology to obtain the original data of CNKI Chinese core journal papers.Data analysis system to analyze the content of the front page the user input based,database-related papers extract the raw data,finally realized the analyzing of corresponding function module.The results show that: the successful implementation of this system improves the researcher analysis efficiency of massive data,the analysis results with higher accuracy,provides a more convenient stabilize and effective way to the researchers analyzed journal articles.