Application of a Modified Holt-Winters Method and of Neural Network Methods to Medium Term Sales Forecasting

Authors:	Kotsialos Apostolos, Technical University of Crete, Greece Papageorgiou Markos, Technical University of Crete, Greece Poulimenos Antonios, Technical University of Crete, Greece
Topic:	5.2 Manufacturing Modelling for Management and Control
Session:	Production Planning and Scheduling - II
Keywords:	Sales forecasts, feedforward neural networks, Holt-Winters method

Abstract

The problem of medium to long term sales forecasting raises a number ofrequirements that must be suitably addressed in the design of the employed forecastingmethods. These include long forecasting horizons (up to 52 periods ahead), a highnumber of quantities to be forecasted, which limits the possibility of human intervention,as well as frequent introduction of new articles (for which no past sales are available forparameter calibration) and withdrawal of running articles. The problem has been tackledby use of a modified Holt-Winters method as well as Feedforward Multilayer NeuralNetworks (FMNN) applied to sales data from two German companies.