SQL Codes to Implement the Bayesian Classification

Na LI, Lin-tong ZHANG, Zhi-gang ZHANG

Abstract


Bayesian classification is a hotspot of machine learning, the purpose of this paper is to train an efficient Bayesian classification based on SQL. It expounded the Bayes' theorem and the concept of Bayes classifier as the foundation, and then, used the recursive algorithm to create a Bayesian data set; the SQL query codes to count classification properties, decision attribute values and classcondition in the training sample, and to calculate the corresponding probability; and the SQL query - update codes to mark the most optimal decision value, thus, the high accuracy of Bayesian classifier model had been trained. Further, we applied the above classification system to predict and determinate the students' grades. It is preliminarily shown that Bayesian classification method has better application prospect in predicting students’ grades.

Keywords


Bayesian classification, Recursive algorithm, SQL


DOI
10.12783/dtcse/mmsta2017/19668

Refbacks

  • There are currently no refbacks.