Demand forecasting in a company : a case study from footwear industry

• Autor(a) principal: Pinto, Maria Francisca de Lima Teixeira
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.14/42549
• topic: Demand forecasting; Footwear; Time series; Regression; Artificial intelligence; Previsão da procura; Calçado; Séries temporais; Regressão; Inteligência artificial; Domínio/Área Científica::Ciências Sociais::Economia e Gestão

Demand forecasting has been investigated for decades, in several areas, such as manufacturing, logistics, and finance, due to its importance in corporate planning and decision-making. Several methods have been tested in different industries, but there is still no consensus among authors, as to which method should be regularly applied since market characteristics differ from company to company. The purpose of this study is to identify the demand forecasting method with the highest accuracy for the characteristics of the data provided by the Portuguese footwear company 8000Kicks, and the reasons for this method have better results than the others tested. A quantitative study is carried out, in the form of problem-solving. The aim of this research is to help solve the company’s problem of lack of efficiency in the use of company resources, impacting its planning and decision-making. Time Series, Regression, and Artificial Intelligence models were selected and tested, to analyse their accuracy, according to the chosen performance measure, Mean Square Error (MSE). The Artificial Neural Network model revealed better accuracy, with the lowest MSE of the models tested, with a test value of 8,5865E-06, followed by Nonlinear Regression. It is concluded that, for this study, the nonlinear models appear to have better results when compared to the linear models, due to their characteristics of adaptability, better fit to the data, and ability to capture complex relationships and dynamic processes.