##plugins.themes.huaf_theme.article.main##
Abstract
The study focuses on presenting the results of mining the data set which have collected and stored since September 2017 about the results getting MOS Excel certificates of students at the University of Economics, Hue University. The research results have high practical significance for students, lecturers with relevant expertise, and training managers of the University of Economics in the trend of improving investment quality, approaching international standards in the trend of globalization. In addition, the study provide a lesson for educational institutions that are starting to approach and pay attention to this issue. To achieve the goal of the study, we use the open source data mining software suite WEKA (Waikato Environment for Knowledge Analysis) with the application of an attribute extraction algorithm using the CfsSubsetEval function to find out. the important attributes stored in the accumulation process that have an influence on predicting the cumulative results of students' MOS Excel. At the same time, we apply the classification technique based on the decision tree algorithms to both data sets before and after extraction to evaluate the obtained results. The results show that the attributes time to study, time to learn directly, the total number of times to participate in the mock test, and the average result of the mock test are four important attributes that are retained after extraction. Finally, the classification model based on the J48 decision tree algorithm is the best model proposed to predict cumulative results with an accuracy rate of above 77%.