Font Size: a A A

College Instructor Academic Warning Management Mode And Its Application Based On Kernel Principal Component Analysis

Posted on:2016-07-26Degree:MasterType:Thesis
Country:ChinaCandidate:J L ZhengFull Text:PDF
GTID:2297330476954287Subject:Mathematics
Abstract/Summary:PDF Full Text Request
Academic warning is an important measure to strengthen the style construction and quality control of university education, currently academic warning system in Chinese universities has not thoroughly implemented. Academic research on early warning analysis is also only stating in deductive, inductive method of qualitative research methods, lessons learned, historical research, comparative study, these methods can not quantitatively reflecting the contribution of each index size, inaccurate, errors and other shortcomings. In view of the above situation, the main work is as follows:Firstly, this paper discussed the role and necessity of basic connotation and early warning system in higher student, investigated the standard of warning system in different levels. Secondly, this dissertation also set principles for established index system of early warning indicator system studies. Thirdly, according to the index system, this dissertation discussed data collection and data preprocessing. Fourthly, although the PCA to extract its strong capability in a wide range of applications, screening indicator shows its advantages, it was not a good solution to bring strong correlation index nonlinear problems discussed in this analysis KPCA in the Treatment of nonlinearity and would not show strong advantage correlation. Fifthly, this dissertation established warning model which was based on KPCA of college students’ academic warning model for counselor perspective, therefore determined the division of early warning and warning level range. It also provided an intuitive quantitative basis, using MATLAB programming to analysis data.Finally, this dissertation used sample from 59 students in Hebei Normal University of Science and Technology in 2013. In addition, by using primary data simulation results, researcher showed that the results of studies warning model is better than the traditional method. The main reason is fixing up the lack of qualitative research methods, effectively enhance early warning targeted, therefore improving operability and effectiveness of the college counselors scientific studies. Figure2; Table18; Reference 61...
Keywords/Search Tags:academic warning, kernel function, principal component analysis, kernel principal component analysis
PDF Full Text Request
Related items