Applied Time Series Analysis - STQM 4020

This course focuses on time series analysis, modeling and forecasting, with emphasis on practical applications in business and other areas. Applied Time Series Analysis are performed using the software R for most of the computational exercises. Upon completion of the course, the students will be able to carry out basic Time Series analysis and fit a model to data.

Offered during the following semester: Fall 2022

List of Tentative Topics:

1. Introduction to Time Series Data Analysis
• Administrative matters • Course overview • Intro to R and R Studio

2. Trend and Seasonality
• Working with Seasonal Relatives • Remove and incorporate seasonality in time series • Stochastic vs Deterministic trends • Stationarity Tests (Mann Kendall, Spearman, Augmented Dickey Fuller)

3. Autocovariance and Autocorrelation Functions
• Learning theory through applied examples • Outliers detection • Handling missing data

4. Box & Jenkins Models - Autoregressive (AR) and Moving Average (MA) Processes
• Intro to Traditional Box-Jenkins Models • ARMA and ARIMA Processes • Seasonal ARIMA Models • Periodic ARMA Models

5. Time Series Modeling and Forecasting: Fitting Models to Data
• ARIMA Models in R • Model Performance Metrics • Intro to Forecasting, Averaging Techniques

6. Model Identification and Parameter Estimation: Model Diagnostics and Selection
• Model Diagnostics, Residual Analysis and Model Selection • Parameter Estimation • Applied Computing Exercises using R

7. State-Space Models and an Introduction to Bayesian Statistics
• State-space Models, Bayesian Statistics • Dynamic Linear Models with Application

8. Other Topics
• Scenario Generation • Multi-model ensembles