Research on Remanufacturing Spare Parts Demand Forecast Based on ARIMA and LSTM

With the increasing number of vehicles worldwide and stricter environmental regulations, the automotive remanufacturing industry has been presented with huge development opportunities. In the Chinese market, as the government actively promotes a circular economy and consumers' environmental awareness is increasing, the automotive remanufacturing industry is also expected to experience a rapid growth period. How to improve the efficiency of related business operations and optimize inventory management capabilities in a constantly changing market environment has become a focus for remanufacturing enterprises. This paper takes B Company as the research object and, through the collection and processing of large amounts of data, finds that it lacks a scientific sales forecasting method, leading to an unreasonable inventory storage of remanufactured parts. By comparing the fitting effects and forecasting errors of the ARIMA (Differential Moving Average Autoregressive Integrated Model) and LSTM (Long-Short Term Memory) neural network two prediction models, the paper ultimately determines that LSTM is more suitable for B Company's sales forecasting. The results show that the research has certain guidance and inspiration for other remanufactured parts enterprises.