An Improved Weighted Naive Bayesian Classification Algorithm Based on Multivariable Linear Regression Model
Xingang Wang, Xiu Juan Sun · 2016
Bias classification is a overuse and efficient classification method in data analysis, but attribute independence assumption affects its performance[1]. In view of these issues, this paper proposes a weighted naive Bayesian algorithm based on MLRM (multiple linear regression model). First, through MLRM to analyze the correlation between attributes, then, this correlation value as the weight coefficient, Finally, using the WNBC (weighted naive Bias classification algorithm) to classify. The experimental results show, that the improved algorithm can down the classification consumption time and improve the classification accuracy rate, and effectively improve the performance of the NBC (naive Bayesian classification algorithm).