Year of Graduation
Data Analysis Applying in Middle School for Current and Forward Education Path Recommendations
This paper introduces and proves different abilities of applying data analysis technics in middle school blended learning environment. Author starts from discussion of various technics, which tend to be useful in educational data analysis and classifying students’ preferences and talents. The main research aim is to create and prove recommendation models over data which is collected in modern middle school, based on logistic regression with l1-normalization, and 3 types of Naïve Bayes methods. This data contains grades, participating in students’ competitions and extracurricular activities and helps to choose current and forward education path. Recommendation system over basic and advanced courses showed 97% of top5 accuracy (better than random model with p-value=0.001), system over specialty showed 82% of top3 accuracy (p-value=0,01). In the conclusion of the paper author describes system, where blended learning and educational data analysis are combined. This system is supposed to be an example of new learning management system’s generation.