Seleción del modelo estadístico adecuado para el manejo de la fluctuación de la demanda en PYMES

Abstract

Enterprises should generate demand forecasts to estimate products sales in a future period, thus, statistical modeling is an indispensable instrument for it. The main objective of this essay is to generate an algorithm, which can select the suitable statistic model to manage demand fluctuation. The algorithm was based on historic data from small and medium enterprises (SMES), to give support in the decision-making process. Causal statistical methods were used such as linear, exponential and quadratic regression, also, time series methods were used regarding moving averages and exponential smoothing. Historic demand data from a commercial enterprise in Cuenca was analyzed with the proposed methods through an engineering statistic software. The purpose was to obtain the statistical technique with the least of errors along with other statistical parameters that helped to determine the analyzed values quality. The adequate model is the one that acquired the least of errors in its projections, which are directly influenced by the statistical normality and homoscedasticity of the data series.