Research On The Application Of Data Mining Technology In The Analysis And Prediction Of Achievements In Colleges And Universities

With the increase of college students at present,universities educational management system has accumulated a large amount of data.But the utilization of those data are mainly used in simple query and statistic.Those data related to the academic performance and the performance of the national college English test band 4 have not been fully excavated.Data mining is the process of extracting unknown,but valuable information from large amount of incomplete data.In this paper,in order to understand the main factors influencing the students' English achievement and master students' academic status in time are of great significance to improve teaching quality and personnel training.This paper studies the K-neighborhood classification algorithm,analyzes the influence that neighbor distance has on classification,and introduces Gaussian weight function,which gives different voting weights according to the distance.So the weighted K-nearest neighbor classification algorithm is constructed.To solve the problem that the classification efficiency of the K-nearest neighbor algorithm is low when the data set is large.This paper further improves the algorithm and establishes the partition-weighted K-neighbor algorithm.The K-nearest neighbor algorithm and the partition-weighted K-nearest neighbor algorithm are applied to the classification and prediction of the results of the CET-4 test.The statistical factors are used to study the related factors affecting the results of the CET-4 test,and the input features and neighboring numbers in the classification forecast are selected.By comparison,although the classification effect of partition-weighted K-nearest neighbor algorithm is not significantly improved,its classification stability is better than that of K-nearest neighbor,and its classification time is reduced from 11.39 seconds to 5.22 seconds,the classification efficiency is increased by 118%.In order to find out students with potential risks of abnormal graduation,and promptly warn and supervise them,this paper establishes the model of decision-tree classification.It can effectively predict the students with abnormal graduation through inputting test scores,or the number of make-up exams and restudies.This paper has 31 figures,12 tables and 52 references.