Research on Predicting Students’ Performance Based on Machine Learning
Download as PDF
Liu Ruochen, Mei Wenjuan, and Liu Jun
Machine learning is one of the most core and hot technology of artificial intelligence at present. It can automatically identify patterns and discover rules based on a large amount of data, predict students' learning performance, and provide possibilities for more reasonable teaching evaluation and personalized learning. Taking the final mathematics scores of students in two Portuguese schools in the medium education as an example, this paper analyzes the characteristics of students' stage scores, personal personality, social relations and daily performance. After dimensionality reduction and other preprocessing of data sets by PCA and other methods, the final mathematics scores of students in one academic year are classified and predicted by SVM and decision tree algorithm respectively, and relevant factors affecting students' scores were analyzed. Finally, it concludes that schools can focus on students' family status, bad habit and ordinary grades to enable students to perform better.
Teaching quality, support vector regression, decision tree