Implementasi Data Mining untuk Klasterisasi dan Prediksi Kelompok Keluarga

Abstract

tahap selanjutnya, sedangkan metode decision tree digunakan sebagai algoritma prediksinya. Hasil penelitian menunjukkan bahwa kedua metode ini berhasil menyelesaikan masalah dengan tingkat akurasi yang sangat tinggi yaitu sebesar 95,3%, presisi sebesar 95,4%, dan recall sebesar 95,3%. Abstract The grouping of families into rich and poor clusters is very much needed as a reference for various future activities such as government assistance or other related parties. Data mining is one approach that can be used to solve this problem. Data mining methods that are suitable are clustering and prediction. This study aims to implement data mining for clustering and predicting family groups. There are two algorithms used in this study, namely kModes and decision tree. The kModes algorithm functions to generate clusters that will be used in the next stage, while the decision tree method is used as the prediction algorithm. The results showed that these two methods succeeded in solving the problem with a very high level of accuracy, namely 95.3%, precision 95.4%, and recall of 95.3%.