Sèries temporals i alguns exemples

Abstract

[en] Any phenomenon that evolves in time, such as the average daily temperature from a region, or monthly sales of a company, from a probabilistic point of view, can be attempted to model using a time series. During this project, we focus on those series that can be decomposed classically, explaining the main characteristics of the most important classic models associated with them. To illustrate the theoretical results, we start from three different sets of observations: the average daily temperature of Madrid, the monthly sales of a company and the number of monthly admissions to ICU (Intensive Care Unit) of a hospital. This way, in parallel with the theoretical concepts explained, we are trying to adjust a probabilistic model suitable for each of the examples, in order to explain accurately how each phenomena behaves. Once justified, that the models found are correct using certain diagnostics, our proposal is to predict how each set of data will evolve in the future.