Supplier Score Prediction Using Hybrid Neural Network Model Based on Simple Exponential Smoothing

摘要

This paper proposes a hybrid model combining artificial neural networks (ANN) and simple average exponential smoothing (SES) forecasting models, termed as the ANNSES model. The proposed model attempts to incorporate the linear characteristics of SES and nonlinear patterns of ANN for predicting the score of suppliers in an e-procurement system of an automobile industry. The MAPE and RMSE errors obtained indicate that predictions upto a month ahead was accurate using the hybrid model compared to those obtained using ANN and SES forecasting models individually.