Enhancing Time Series Forecasting with the Advanced Cumulative Weighted Moving Average Technique

Abstract

Accurate forecasting is applied to several industries, especially in automotive engineering, where tasks are prediction of car spare parts demands and estimation of maintenance works to optimize inventory levels and reduce costs. Traditional methods for forecasting, Moving Averages, and Exponential Smoothing often need to capture the dynamic nature of these demands. This paper introduces a new technique in the family of weighted moving average techniques: the cumulative weighted average method. It follows that the more recent data points will be given progressively increasing weights, making this type of forecast more accurate. This is again illustrated by a numerical example in which we compare CWMA against traditional methods and show that CWMA produces more accurate forecasts, as evidenced by its lower Mean Squared Error. The study further envisages the potential of CWMA to enhance forecasting quality in the automotive environment, specifically in managing spare parts inventory or maintenance planning activities. It is recommended to validate using real-world data and further research into possible improvements to enhance the accuracy and applicability of CWMA.