Validation and Selection between Machine Learning Technique and Traditional Methods to Reduce Bullwhip Effects: a Data Mining Approach

摘要: The aim of this paper is to present a methodology in three steps forecast supply chain demand. In first step, various data mining techniques are applied in order prepare data for entering into forecasting models. second the modeling an artificial neural network and support vector machine presented after defining Mean Absolute Percentage Error index measuring error. structure artificial selected based on previous researchers' results article accuracy of network increased by using sensitivity analysis. best forecast for classical methods (Moving Average, Exponential Smoothing, Exponential Smoothing with Trend) resulted based on prepared compared result support vector proposed network. results show that can more precisely in comparison other methods. Finally, methods' stability analyzed raw even effectiveness of clustering analysis measured.