Demand forecasting by artificial neural networks in petrochemical industry : A case study on Petkim petrochemical corporation

Artificial Neural Networks which have an important role in artificial intelligence studies are inspired by the structure and functional aspects of biological neural networks. ANN are usually used to model complex relationships between inputs and outputs. This model has been applied to the problems which are difficult and complex in diffirent fields and successful results have been obtained. This method which has a wide field of applications has been used for demand forecasting area in recent years.In this study, it is aimed to forecast PVC ( Polyvinyl Chloride ) plastic raw material demand in Petkim which is the sole petrochemical products manufacturer in Turkey by the artificial neural network method. Demand forecasting model with ANN has been built by using the data in the period of January 2005 ? December 2011. ANN model has been tested and the results have been evaluated. Additionally, demand forecasting model has been developed by multiple linear regression analysis which is a conventional forecasting technique. Then, forecasting performance of two methods have been compared. The results have demonstrated that ANN model has provide more accurate prediction. In conclusion, it has been determined that ANN is an efficient tool for forecasting.Keywords: Artificial Neural Networks, Multiple Linear Regression Analysis, Demand Forecasting, PVC( Polyvinyl Chloride ) Plastic Raw Material.