Spare part supplier selection model using decision tree classification techniques: J48 Algorithm

Abstract

Spare parts are goods that consist of several components that form a single unit and have certain functions. To facilitate companies in selecting suppliers, a supplier selection model is needed to make it easier for companies to select suppliers and make it easier to see strategic directions to take several criteria from suppliers to achieve priorities. This research was conducted at a manufacturing company engaged in car spare parts with rubber raw materials. In this study the problems that occur are the difficulty of the company in selecting suppliers based on criteria that are in accordance with the company and the delay in the receipt of raw materials, and the lack of raw materials supplied from each supplier. This study aims to classify suppliers based on desired criteria by the company and to design a supplier selection model in the long term. The J48 algorithm is produced by a supplier selection model with 2 rule selection models to classify efficient suppliers and inefficient suppliers. The accuracy of the Decision Tree model is 90.8547%, the MAE error value is 0.1256. From the J48 algorithm the biggest gain is the criteria for quality, price, delivery, and warranty and complaint services.