A Model for Forecasting the Product Demand in Tile and Ceramic Industry

Document Type : Original Article


1 Associate Professor, Shahid Beheshti University.

2 MA, Semnan University

3 MA, Tarbiat Modares University.


Forecasting the product demand is one of the most important activities of every organization for planning the sale and finally making a comprehensive plan. It in fact determines the amount of the activities an organization carries out in the future and also provides the managers with a clear understanding of the amount and quality of the respected activities. In this article, a model based on Artificial Intelligence and Data Mining algorithms is proposed for predicting the amount of sale in tile and ceramic industry. The proposed model is a hybrid model consisting of dimensionality reduction, clustering and forecasting. In order to construct the model, this research uses independent component analysis, Manifold learning, K-means clustering and support vector regression. The present research studies 50 cases of Irana Tile Company's past 3 years of monthly sales. Because of the decrease in generic and sampling errors, the results obtained from this model show a superior precision compared to other traditional forecasting methods.