APPLICATION OF DATA MINING TOOLS FOR ANALYSIS AND FORECASTING OF LEARNING SATISFACTION

UDC 004.89
DOI:10.26102/2310-6018/2019.27.4.006

P.B. Skripko, V.S. Dunin, E.A. Kleimenov


The article discusses the features and results of using data mining tools, an application for building See5 decision trees, with which a number of parameters are determined on the basis of sociological survey data that are most important for assessing student satisfaction. Along with the features of using a software tool for data analysis in the work, the procedure for preparing data for downloading and further analysis is shown. At the same time, not only file formats loaded into the application are presented, but also approaches to cleaning and aggregating source data, including in order to reduce their dimensionality. Interpretation of the results of constructing decision trees – classification rules obtained during processing, is aimed, inter alia, at choosing the optimal settings for processing the source data. The classifier constructed using the data of a sociological survey showed that such parameters as the atmosphere of interpersonal relations, recognition of the achievements of the student, and the prestige of the educational organization have a significant impact on the classification. The paper also describes the procedure for assessing the predicted values of student satisfaction based on the input of the estimated values of the classification features.

Keywords: data mining tools, decision trees, decision rules, classifiers, forecasting student satisfaction.

Full text:
SkripkoSoavtors_4_19_1.pdf