Weight Optimization of Weighted Naive Bayes Classifier for Efficient Classification

The Weighted Naive Bayes (WNB) classifier is an effective classification method based on the Naive Bayes algorithm; however, determining the appropriate weights within this algorithm remains a significant challenge. The number of optimization based WNB applications in the existing literature is quite limited. Grid Search methods used for weight determination often suffer from high computational costs and an inability to reach the global optimum. Although the Fast Weighted Naive Bayes classifier offers faster solutions, it remains limited in terms of classification performance. Therefore, optimizing weights is of great importance in achieving both computational efficiency and high classification accuracy. In this study, Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) methods were employed to optimize the weights of the WNB algorithm. The proposed GAW-NB and PSOW-NB models were evaluated on five different datasets using 5-fold cross-validation. The experimental results demonstrated that both methods achieved significant improvements in terms of both execution time and classification performance.