PERFORMANCE ANALYSIS AND COMPARISON OF DEMAND FORECASTING MODELS AT AN ELECTRONICS COMPANY

Tóm tắt

Time series forecasting encompasses both conventional statistical methods such as Moving Average, Exponentially Weighted Moving Average, and AutoRegressive Integrated Moving Average and contemporary machine learning techniques, including Artificial Neural Networks and Long Short-Term Memory. Each technique possesses inherent strengths and limitations, making the selection of an appropriate method dependent on the industry and data characteristics. This study investigates demand forecasting in the electronics sector, specifically for electronic components, to optimise resource allocation in a cost-effective and efficient manner. By evaluating the performance of various forecasting methods, comparing their effectiveness, and identifying the approach with the lowest error and highest optimisation potential, this research contributes empirical insights that support informed decision-making and promote operational efficiency.