Sèries temporals amb memòria llarga: models ARFIMA

Abstract

[en] Time series analysis plays a vital role in understanding various phenomena in a wide variety of fields. Over the years, more complex statistical models have been developed that allow us to analyze time series in a more accurate way. This project studies in depth one of them: the ARFIMA models. These models are very useful for examining long memory time series. The main purpose of this project is to provide a comprehensive analysis of these models, including their theoretical foundations, estimation methods and a practical implementation. In particular, the $R$ programming language is used to evaluate the possibility of modeling the time series that shows the air quality of Barcelona between 2017 and 2019 using an ARFIMA model.