The Application Of Data-mining To The Student Psychological Management System

Good mental health is vital to college and university students. As theiraggravating psychological problems have become a big concern for the administrativeauthorities of higher education, the psychological counseling centers of most collegesand universities come to adopt the psychological management system to conductpsychological tests, statisticizing as well as establish and keep records of mental heathfor their students, with a view to achieve instant statisticizing, timely feedback andtimely decision-making on prevention and intervention by substituting computers forpsychological counselors to a certain extent.The paper aims to embed data-mining techniques into the existing psychologicalmanagement system, discussing in detail the design and achievement of the modulesof data-mining techniques appropriate for psychological management system. Itmakes an introduction to the vital decision-making algorithmic techniques and thelinear fitting employed in the course of development and addresses the application ofthe decision-tree algorithm to analyzing the factors affecting the mental health ofcollege students. The major tasks the paper intends to accomplish include:(1) togenerate a decision tree through C4.5algorithm based on the existing psychologicalmanagement system, the available psychological test results and the essentialinformation of the students. A new approach to pruning the decision tree will beadopted, which will mask off the derived rules of low confidence level and thereforemake derived rules more efficient and accurate, rendering decision-making morefruitful.(2) Linear fitting is employed to update the test results of various measuringscales, which can screen the individuals with psychological abnormalities with thegreatest accuracy possible, thus laying a sound basis for psychological counselors andrendering the existing student psychological management system more useful andfunctional and more satisfactory to both counselors and students.The paper attempts to embed data-mining techniques into the existing psychological management system, ensuring convenient, swift and accurate detectionof the correlations among various factors from a mine of data and producing results ofgreater reference value without affecting adversely the other functions of the system.In the process of development and embedment, appropriate data-mining techniqueswill be selected and the unparalleled solutions such techniques offer will be presented.On the other hand, an overall grasp of the student psychological management systemcan be achieved in the course of design, development and application, making itpossible to construct a scientific service system or pattern for keeping the mentalhealth records of college and university students and enabling the studentpsychological management system to play an efficient and facilitating role in mentalhealth education in colleges and universities.The paper expounds the demands for information analysis based on studentmental health records and employs the decision-tree algorithm to classify the data.Methods such as comparison, contrast, aggregate analysis and theoretical verificationare used to formulate means for data selection, cleaning and conversion. Lastly theapplication and achievement of data-mining techniques are discussed, along with asummary and speculation of the system in question.