Previsão de demanda em indústrias de eletrodomésticos por meio da integração de métodos quantitativos e qualitativos
The constant search for knowledge of the consumer market leads companies to develop mechanisms to analyze information about the intentions of consumers. The current competitive level, imposed by globalized markets, forces companies to develop skills to forecast demand in the short, medium or long te...
| Autor principal: | Ferro, Wílian Assmann |
|---|---|
| Formato: | Dissertação |
| Idioma: | Português |
| Publicado em: |
Universidade Tecnológica Federal do Paraná
2018
|
| Assuntos: | |
| Acesso em linha: |
http://repositorio.utfpr.edu.br/jspui/handle/1/3058 |
| Tags: |
Adicionar Tag
Sem tags, seja o primeiro a adicionar uma tag!
|
| Resumo: |
The constant search for knowledge of the consumer market leads companies to develop mechanisms to analyze information about the intentions of consumers. The current competitive level, imposed by globalized markets, forces companies to develop skills to forecast demand in the short, medium or long term. In this context, the integration of qualitative and quantitative demand forecasting methods with expert adjustments has become one of the main methods used to improve forecast accuracy. Through a structured review of the literature on this subject, it was possible to identify a lack of studies regarding the theme presented with an emphasis on structuring the judgment of specialists. Therefore, as the main objective of this dissertation is to structure these expert judgments way multiple criteria method of decision support. For this, three time series are used, each of which is composed of a family of household appliances; the individual prediction of each series is initially performed with the use of family techniques (ARIMA and Exponential Smoothing). Following the best methods are chosen by AIC criteria and are combined via means arithmetic, weighted, harmonic and geometric. The measure of accuracy used to evaluate the predictions is MAPE. The results obtained only emphasize improvement of the alpha family prediction when this is combined for the other beta and gamma families, this did not occur. Finally, the best quantitative forecasts obtained are adjusted by the experts via structured judgment. With the implementation of the methodology, a reduction of MAPE was obtained, in relation to the best quantitative predictions, of: 17.00% for the 1st family; 10.55% for the 2nd family; 72.93% for the 3rd family and 10.57% for the global forecast. Therefore, it is evident that, with the obtained results, the accuracy of the forecast occurs when the proposed methodology is implemented. |
|---|