Student Achievement Prediction Comparison Of Naïve Bayes And Svm Using Ai Optimization In Smpn 5

Azzahra Permatasari Suwandi (a:1:{s:5:"en_US"s:21:"Universitas Nusaputra"})
Sudin Saepudin (Information system / , Nusa Putra University)
Gina Syabani Yuda (Information system / , Nusa Putra University)

International Journal of Health, Engineering and Technology
Vol. 5 No. 1 (2026): IJHET MAY 2026

This study aims to build a predictive model of student learning achievement by comparing the performance of Naive Bayes algorithm and Support Vector Machine (SVM) optimized using Synthetic Minority Over-sampling Technique (SMOTE) and Grid Search. Methods used in this study include data collection, preprocessing, data sharing, application of SMOTE to handle data imbalances, as well as parameter optimization using Grid Search. Next, the model was built using Naive Bayes algorithm and SVM, then evaluated using accuracy metrics to determine the best performance. The results showed that the accuracy of Naive Bayes algorithm before SMOTE application was 71%, but decreased to 56% after SMOTE application. Meanwhile, the SVM algorithm showed stable results with an accuracy of 68% both before and after the application of SMOTE. This shows that optimization techniques do not always improve the performance of the model, depending on the characteristics of the data used. Thus, SVM models are considered more consistent, while Naive Bayes is more sensitive to data changes. The resulting Model can be used as an aid in decision-making in the field of education to more accurately identify the level of student achievement.