COMPARISON ANALYSIS OF NAÏVE BAYES AND K-NEAREST NEIGHBOR METHODS ON THE PREDICTION OF ACADEMIC POTENTIAL IN SMK TI BALI GLOBAL BADUNG

Abstract

SMK TI Bali Global Badung is a vocational high school that focuses on information technology. SMK TI Bali Global Badung has an academic information system that is still conventional. The potential possessed by students can only be seen in terms of the value obtained, but supporting factors such as place of residence and physical condition can affect the potential of these students. Along with the development of information technology, data mining can provide solutions for schools to find academic potential based on stored data. This data mining system will summarise data from various data and information by analyzing specific patterns or relationships from several data types. Based on the current research results, this study will use the method to predict the academic potential of SMK TI Global Badung students, namely Naive Bayes and K-Nearest Neighbor. Prediction model of academic potential of SMK TI Bali Global Badung using Naive Bayes and KNN methods built using criteria and sub-criteria in the calculation process. In this study, seven criteria were used: character, academic activity, participation in intra-curricular activities, participation in extra-curricular activities, place of residence, and socioeconomic status. Based on the tests, Naive Bayes produces an accuracy value of 43.48%, a precision of 63.64, and a recall of 31.11. K-Nearest Neighbor creates an accuracy value of 57.97%, a precision of 62.5, and a recall of 88.89. In further research, it can also be considered to combine methods to increase the effectiveness of the predictions produced.