Research On The Way To Improve The Quality Of Master’s Thesis

Master’s thesis plays an important role in the teaching system of colleges and universities.Its quality not only represents the learning and research results of the three-year postgraduate study,but also represents one of the most important indicators of teaching quality in the school’s master’s training plan.Therefore,in 2014,the Ministry of education stipulated that the master’s thesis should be randomly checked and reviewed every year.Therefore,all schools have strict requirements for master’s thesis,which must be submitted to experts inside and outside the school for examination.Once the review quality of master’s thesis is poor or even unqualified,it will affect the normal graduation of graduate students and the teaching reputation of the school.Taking the master’s thesis of school a as an example,this paper uses data mining method to analyze the external review data of the master’s thesis,finds out the main factors that affect the quality of the master’s thesis,and gives relevant suggestions in the paper quality improvement,hoping to improve the quality of the master’s thesis.In this paper,the Graduate School of school a obtained the evaluation data of master’s thesis in the past three years.After data cleaning,1505 effective evaluation data of master’s thesis in the external audit were obtained.First of all,using descriptive statistical analysis to evaluate the overall quality of the thesis,it is found that the overall quality of the master’s thesis in school a is still good,but the proportion of high-quality thesis is not high enough;in addition,there are differences in the quality of the thesis in various disciplines in the school,from high to low,they are politics,language,economy,management and law.On the basis of the overall analysis of the thesis quality,this paper uses the methods of text mining,such as Chinese word segmentation,removal of stop words and word cloud chart,to analyze the external review opinions of the master’s thesis experts in school A.It is found that the problems of the thesis quality in all disciplines of school a are different,mainly in 10 aspects.The common problems are the lack of theoretical depth and the lack of comprehensive and specific content analysis,andaccording to these,the paper Put forward corresponding modification suggestions.In the risk assessment of the quality of the master’s thesis,the association rule mining and k-nearest neighbor classification model are established from the qualitative and quantitative perspectives,respectively,to predict the risk of the thesis.The association mining model of paper quality risk shows that the research direction of this paper is application-oriented,and the graduate thesis with tutor’s evaluation score less than 70 points has a high risk,which should be monitored emphatically;the prediction of k-nearest neighbor classification model of paper quality risk shows that 85.7% of the unqualified thesis is correctly detected by the risk assessment model,and 84.7% of the qualified thesis is correctly detected.Finally,for the papers that need to be monitored,according to the suggestions of text mining,we can modify them to improve the quality of their papers and reduce the risk of unqualified papers.