Weight Optimization of Weighted Naive Bayes Classifier for Efficient Classification

Authors : Gamzepelin Aksoy, Murat Karabatak

Pages : 51-63

Doi:10.5281/zenodo.15719600

View : 31 | Download : 20

Publication Date : 2025-06-25

Article Type : Research Paper

Abstract :The Weighted Naive Bayes classifier is an efficient classification algorithm based on the Naive Bayes Algorithm. However, the determination of weights in this algorithm is an important problem. By using the Grid search method, the optimum solution is not reached and the algorithm works too slowly. Fast Weighted Naïve Bayes classifier is used to find weights quickly, but the performance of this algorithm is limited. Therefore, optimizing the weights has a great importance in terms of both time and achieving high performance. In this study, Genetic Algorithm and Particle Swarm Optimization methods were used to optimize the weights of the Weighted Naive Bayes classifier. The performance of Genetic Algorithm based Weighted Naive Bayes (GAW-NB) and Particle Swarm Optimization based Weighted Naive Bayes (PSOW-NB) methods were examined on five different sets with 5-fold cross validation testing method. The results of the experiments showed significant results both in terms of speed and classification performance.
Keywords : Ağırlıklı Sade Bayes, Sınıflandırma, Genetik Algoritma, Parçacık Sürü Optimizasyonu.

ORIGINAL ARTICLE URL