Effective Application of Naive Bayesian Classifier for Personal Online Learning Networks

Abstract

Naive Bayesian classifier can be used to classify news and patients, but there are few studies on the classification of educational data. Based on Naive Bayesian algorithm. This paper studies the relationship between course achievement and employment salary. Quantitative method is adopted as research methodology. The sample data sets were collected from Personal Online Learning Networks, which consist of the Student Performance Management System and Student Employment Management System. The sample category labels were constructed and the Hold-Out method was used to divide data sets into training sets and testing sets. 15 courses’ performance as feature vector and employment wage as category, if the attribute condition was independent, a Naive Bayesian Classifier was established. The result indicating the higher the grade of DAWEB, ICT, INT and WNDW courses, the higher the employment wage. The conclusion is in accordance with the actual situation: Four courses mainly train students’ comprehensive practical ability. The students who have stronger practical abilities are highly demanded by employers, hence, the higher salary will be provided. At the end, regarding the class conditional probability of $$P(x_{i} = E|s = H)$$ (Performance = E, salary = H) as the weight of courses, build a topological structure diagram of courses.