Model Prediction Using Random Forest Algorithm and Time Series Analysis Parts Distribution with in Indonesia Automotive Industry

Car spare parts are parts of a car that consist of components that form a certain unity and function, there are some parts that are damaged more often and therefore must be replaced more often. In this research, to predict the availability of spare parts using the times series moving average prediction method with random forest. The dataset used in this research is 19,068 consisting of location, transaction number, transaction date, part number, quantity and odo-km. Distribution of spare parts classified based on recapitulation of spare part distribution is RAE=54.5541% and RSE=63.8888%, distribution of spare parts based on location is RAE=55.5541% and RSE=65.2118%, maintenance based on ODO-KM is RAE=33.0949% and RSE=41.2241 % and distribution quantity RAE=33.749% and RSE=36.2317%. The results of spare part distribution analysis using the Confusion Matrix with probability accuracy are TP Rate=0.903, FP Rate=0.067, Precision=0.911, Recall=0.903, F-Measure=0.893, MCC=0.857, ROC Area=0.983 and PRC Area=0.948. The aim of this research is to determine the factors needed to build a machine learning-based spare parts distribution model, implementation, measurement and prediction in the automotive industry in Indonesia.