Klasifikasi Kebutuhan Sparepart Dengan Algoritma K-Nearest Neighbor Untuk Meningkatkan Penjualan Sparepart

Abstract

Adequate supply of spare parts will be a supporting factor for consumer confidence in the company. The classification method approach can be applied in analyzing data to apply data mining with the classification method for spare parts needs generated by utilizing data testing consisting of 100 record datasets with a ratio of 90% training data (training data) and 10% test data (data testing). . Implementation of the K-Nearest Neighbor algorithm model on test data (data testing) of 100 data objects, obtaining results that show a new insight in the form of classification of low and high level needs based on 2 categories. No is a category of light needs, consisting of 89 data objects, the category Yes is a category of high needs. Performance evaluation and testing using the RapidMiner Sstudio application is able to provide optimal results with the scenarios that are modeled. This algorithm model has an Accuracy value of accuracy: 93.00% +/- 6.40% (micro average: 93.00%).