The Application Study Of Data Mining In The Evaluation Of Impoverished Colleges Students

With the enrollment of higher education in China in recently, the number of impoverished students rise rapidly. The evaluation is the first mission in managing impoverished colleges students. only if impoverished students are identified and classified correctly,we can provide financial aid correctly. With the impoverished students management information systems are used extensively, a large number of data is collected and accumulated. However, the data is still used on the level of copying, searching and counting,described as ‘data abundance, but information indigence ’. Data mining is the technology what can acquire and mine knowledge from mass data, thus we use Data mining Technique to evaluate impoverished students and improve the evaluation work.This article takes the evaluation work for the impoverished students of Zhujiang College, South China Agricultural University as an example. Use c4.5 decision tree algorithm to construct the evaluation model of impoverished students to solve the problem of evaluation working practice for impoverished students. Firstly, it analyzes the difficulties in the evaluation work of impoverished students. Clear about the aim of exploring. Predict the impoverished students with unknown classification by exploring the historical data of classified impoverished students. Then, based on the original data exported from the information system of impoverished students, it analyzes the relationship between data sets and the data availability, thus to form a two-dimension of evaluation data sheet for the impoverished students. Use property-select algorithm of different combination to simplify the property of impoverished students. Process pretreatment for the data by a series of methods including data cleaning, data conversion, discretization and concept hierarchy, it gets the data sets of impoverished students that need to be explored.The paper uses open source system and applys c4.5 algorithm to the evaluation of impoverished students for the construction of decision tree modeling. Set the nodal point with minimum example, which is two-pair of decision tree branch. Select the concentrated 30% data of impoverished students’ data as the classification effect of the training test model according to the evaluation criteria of classification model. Compare the classification results with the existing results at school. The experimental result shows the precision rate and accuracy are both high and the effect of model classification is good. From the analysis by extracting the rules from the constructed model, we find the important criteria and evaluation rules of impoverished students’ classification, thus it proposes effective suggestions to complete the evaluation system of impoverished students and improve the evaluation work.