International Reputable Journal Classification Using Inter-correlated Naïve Bayes Classifier

Naïve Bayes Classifier has weaknesses in feature independence that made some of the features does not have a significant effect on the decision. This article describes the development of an Inter-Correlated method to optimize feature selection on Naïve Bayes classifier. This concept is designed to ignore the limitations of Naïve Bayes. This method is implemented on international reputable journal rank classification. The dataset used in the testing process is obtained from the Scimago Journal and Country Rank website. The dataset consists of 1491 instances with 9 attributes. SJR Best Quartile is selected as a class label that consists of value Q1, Q2, Q3, and Q4. Based on the results of experiments with an optimized Naïve Bayes algorithm, the algorithm obtained an accuracy value of 59.14%. The optimized algorithm yield a better result compared to the accuracy of the Naïve Bayes before optimization of 50.49%. This shows that the developed method is more efficient and increases the accuracy value to 8.65%.