PENGARUH PROSES DISKRITISASI FAYYAD IRANI PADA AKURASI DATA MINING KLASIFIKASI NAÏVE BAYES

Data Mining Classification is a technique to classify data into a category or class. The classification process can have varying accuracy depending on the compatibility between the data and the algorithm used. One way to improve the accuracy of the classification process is to use data processing. Fayyad-Irani discretization is a pre-processing of data that converts numerical values to categorical. Fayyad-Irani's discretization has been proven to increase the accuracy of classifications using decision tree algorithms. Unlike the decision tree algorithm, the Naive Bayes algorithm is a Bayesian probability based algorithm so that it can have a different effect on the accuracy and timing of the classification. Based on the experiments conducted in this study, it is known that Fayyad-Irani Discretization has succeeded in increasing the accuracy of the Naive Bayes classification process significantly from only 29.31% to 93.26%, but with the processing time needed to be slower with time difference as much as 25.5 minutes slower than the process without discretization. Keywords: Classification Data Mining, Naive Bayes, Fayyad-Irani’s Discretization