Análise da aplicação do modelo ARIMA: um estudo de caso na indústria de reposição automotiva

Abstract

Inventory planning is a challenge for industries, especially for the automotive aftermarket industry, which faces high volatility in market demand levels. In this sense, the present work sought to analyze the application of the ARIMA method in the demand forecast of an aftermarket product of a multinational automotive industry. It was briefly contextualized the current scenario at the company for demand forecasting, determined a representative aftermarket product for the analysis and collected real sales historical data. From an exploratory-descriptive case study, a demand forecast was performed using the ARIMA method applied according to the steps proposed by Box-Jenkins. The seasonal ARIMA (2,1,1)(0,1,1)12 was the best fit model for the time series under study according to AICc and BIC information criteria, and residue analysis. Thus, it was found that the defined model met the method criteria but requires attention as to its accuracy due to a mean absolute error of 59%.