8221 - TIME SERIES ANALYSIS OF ECONOMIC-FINANCIAL DATA
GM-LS - MM-LS - OSI-LS - AFC-LS - CLAPI-LS - CLEFIN-LS - CLELI-LS - CLEACC-LS - DES-LS - CLEMIT-LS - CLG-LS
Course taught in English
Go to class group/s: 31
The analysis of dynamic phenomena is of main relevance in economic and financial studies. The aim of the course is to provide a knowledge of the classical statistical procedures for time series analysis, but also of more modern techniques, based on dynamic linear models (or state-space models). The course intend to provide a solid methodological background and data-analysis skill, with lectures in the computer room and individual and team work. The software will be SPSS and/or R for the first part of the course, and R for the second part of the course.
Part I. Classical analysis of univariate time series
- Descriptive techniques. Decomposition of a time series; trend, seasonality.
- Stochastic models. Stationary processes. Models in the time domain; ARMA and ARIMA models.
- Exponential smoothing. Forecast and model comparison.
Parte II. Dynamic linear models for time series analysis.
-
Dynamic linear models for univariate and multivariate time series. Estimation and forecasting. Kalman filter. Bayesian approach.
-
Extensions. Maximum likelihood estimation of unknown parameters. Unknown covariance matrices and Bayesian inference.
-
Model specification. Trend and seasonal components. Regression components.
-
Advanced topics (Models for series with structural breaks. Stochastic volatility models. Financial application).
The exam consists of an individual work on the analysis of real data and an oral test.
-
C. CHATFIELD, The Analysis of Time Series, Chapman & Hall/ CRC, 2004, 6th ed.
-
G. PETRIS, S. PETRONE, P. CAMPAGNOLI, et al., Dynamic Linear Models with R, New York, Springer (in progress).
R free software is available at http://www.cran.r-project.org/