Author(s): Blanco, Felipe Orestes Cuan Suárez
Date: 2018
Persistent ID: http://hdl.handle.net/10362/40987
Origin: Repositório Institucional da UNL
Subject(s): Car Sales; ARIMA; Hierarchical Forecasting; Regression with ARIMA Errors
Author(s): Blanco, Felipe Orestes Cuan Suárez
Date: 2018
Persistent ID: http://hdl.handle.net/10362/40987
Origin: Repositório Institucional da UNL
Subject(s): Car Sales; ARIMA; Hierarchical Forecasting; Regression with ARIMA Errors
Dissertation presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Knowledge Management and Business Intelligence
This study aims to contribute to a better understanding of the Portuguese automotive market, its behaviour and, additionally, attempt to create better sales prediction models, using linear regression models with ARIMA errors. Through time series analysis, it will be studied how external factors such as socio-economic variables, consumer behaviour as well as market specific factors (e.g. gas prices, car rental services) affects the development of the automotive market. Furthermore, it will be tested if methodologies such as hierarchical forecasting on a bottom-up approach will bring better results compared to forecasting the market as a whole. This understanding and forecasting techniques, if proven successful, can bring higher competitive advantage to companies in this sector, through more accurate planning and strategic management.