620-382 Time Series and Forecasting

Note

Students may only gain credit for one of 620-382 and 619-360 (1997 Handbook).

Credit Points

Coordinator

Prerequisites

620-202 (1997 Handbook 619-202) or 620-371 (1997 Handbook 619-330).

Semester

Contact

Subject Description

This subject introduces the concept of a time series, simple models for time series and the basic principles of time series analysis in the time domain and in the frequency domain. Students learn to use standard forecasting techniques, to carry out simple time series analyses in the time domain and in the frequency domain and to use the computer as an aid to efficient analysis of time series. This subject demonstrates the patterns inherent in real-world time series and the importance of the underlying mathematical theory of time series.

Patterns in time series. Simple methods for exploratory data analysis: smoothing techniques. Smoothing and decomposition. Trends and seasonal variation. Simple forecasting methods: mean, last value. Accuracy of forecasts. Exponential methods. Models for time series: stationarity, autocorrelation; ARMA processes; ARIMA processes. Estimation and model fitting. Spectral analysis. Smoothing by Fourier series. Linear filtering.

Assessment

Up to 3 hours end-of semester written examination; up to 50 pages of assignments may be assessed.