Análise de séries temporais com dados reais e desenvolvimento de modelos de previsão

Abstract

This Undergraduate Thesis (TCC) addresses the development of time series forecasting models, specifically applied to five market variables: Tires, Steel Sheets, Rubber, Electric Energy in Brazil, and Polyethylene Packaging. The study employs a combination of exponential smoothing and ARIMA models, recognized for their robustness in analyzing temporal data with trends and seasonal patterns. The goal is to provide forecasts up to 18 months ahead, balancing medium- and long-term behavioral nuances of the series. The research focuses on optimizing forecasting parameters, aiming for a compromise between sensitivity to current market conditions and the incorporation of historical trends. The duality of methods allows addressing time series with different degrees of volatility and periodicity, providing a deeper understanding and broader applicability. The practical implications of this study are particularly relevant for strategic decision-making in the commercial and industrial sectors, given the critical role these products play in the economy.