Demand Forecasting of Spare-Parts Using the Data Mining Techniques

Abstract

Nowadays, many businesses use traditional inventory management techniques. Proper inventory management techniques save businesses from additional costs. Make traditional and experience-based forecasting is not enough to manage inventory. Collecting, storing, and interpreting data is vital for the future of businesses. Vehicle maintenance is necessary for safe, comfortable, and accessible public transportation services. The quality of the vehicles is crucial for the satisfaction of passengers. Spare parts that are out of stock cause delays in the maintenance of the vehicles and the inability to provide service. On the other hand, spare parts purchased more than needed cause increasing holding costs and unnecessary expenses. The increase in spare part prices after the pandemic affects public transportation negatively, especially in our country. That is why demand forecasting is much more important today. The majority of public transport vehicles are buses. Filters are one of the most common spare parts affecting the operation of the buses among the spare parts. There are so many filter types in the business where the study was conducted. The aim of the study is to minimize the loss with the proposed inventory management practices and to ensure that the maintenance of the buses is done at the right time. In the study, the needs of filters were calculated using “Linear Regression”, “Support Vector Regression”, “Neural Network” and “Random Forest”. The accuracy of the techniques used was proven by using real data.