USE OF FORECASTING TECHNIQUES TO ESTIMATE DEMAND IN SMALL AND MEDIUM-SIZED COMPANIES IN THE TEXTILE SECTOR

Abstract

The aim of this article is to present a comparative analysis of statistical techniques to forecast the demand of clothing. Techniques were validated in a textile company producing knitted garments located in the south of the Mexican state of Guanajuato. A time series was analyzed to identify cyclicity, trends, seasonality, and random variations. Moving average, weighted moving average, exponential smoothing, Holt's method, and Winter’s method were applied to demand forecasting. Model performance was tested by mean absolute deviation (MAD), mean squared error (MSE), and the R2 determination coefficient.The results proved simple exponential smoothing to be the best performing technique for demand forecasting. Forecasting techniques presented as spreadsheets will help decision makers to anticipate their clients’ demand, respond promptly, and improve organizational aspects.The present study attempts to highlight the importance of providing small and medium-sized enterprises in the textile sector with engineering tools to improve their production and business processes by making informed decisions to enhance their competitive advantage.