Fundamentals of Time Series Analysis - Online Course

We will begin with the simple case of a univariate stationary time series. We will discuss the theory of how to recognize autoregressive moving average (ARMA) processes, as well as the intuition behind estimation of these models via maximum likelihood. We will pay special attention to making sure the underlying assumptions of our ARMA models are satisfied in practice, and then use these models to generate forecasts. Once we have a solid foundation in stationary models, we will discuss how to recognize and address non-stationary processes. We will then move onto multivariate times series models and address causality.

Along the way we will discuss practical issues, including how to pick the order of lags or the polynomial order for the trend. Each day will include plenty of hands-on practice, so you will leave with both a firm grasp of the theoretical underpinnings of time series methods and a clear understanding of how to apply them to your own work.

In the blog post Time Series Modeling: Why ARIMA Models Beat Linear Regression for Temporal Data, the limits of regression are compared with the strengths of ARIMA models for improving forecast accuracy. This seminar builds on those ideas with hands-on practice fitting and interpreting your own models.

This seminar will use R for the empirical examples and exercises. To participate in the hands-on exercises, you are strongly encouraged to use a computer with the most recent version of R and RStudio installed. RStudio is a front-end for R that makes it easier to work with. This software is free and available for Windows, Mac, and Linux platforms.

Basic familiarity with R is highly desirable, but even novice R coders should be able to follow the presentation and do the exercises.

If you’d like to take this course but are concerned that you don’t know enough R, there are excellent on-line resources for learning the basics. Here are our recommendations.

This course will be useful for those who want to apply time series methods to their data in an academic setting or at their workplace. It is strongly recommended that participants be familiar with linear regression. No prior experience with linear algebra is necessary.

Day 1

1. Mathematical tools for time series analysis (brief overview)
   1. Difference equations
   2. Lag operators
   3. Higher order difference equations
2. Modeling and estimation for univariate stationary time series models
   1. Intuition
   2. Autoregressive moving average processes
   3. Estimation techniques
3. Estimation Diagnostic checking
   1. Analysis of residuals
   2. Formal statistical tests
4.  Coding
   1. Times series plots
   2. Correlograms
   3. Ordinary least-squares
   4. Maximum likelihood estimation
   5. Diagnostic tests

Day 2

1. Forecasting
   1. In-sample forecasts
   2. Out-of-sample forecasts
   3. Point, interval and density forecasts
2. Multivariate stationary time series models
   1. Cross-correlation function
   2. Estimation
3.  Coding
   1. In-sample forecasts
   2. Out-of-sample forecasts
   3. Point, interval and density forecasts
   4. Cross-correlation function
   5. Multivariate time series estimation

Day 3

1. Nonstationary times series models
   1. Deterministic trends
   2. Stochastic trends
   3. Unit root testing
2. Structural breaks
   1. Theory
   2. Estimation
   3. Interrupted time series
3. Multivariate nonstationary time series models (if time permits)
   1. Spurious regression
   2. Vector Autoregression
   3. Vector autoregression
4. Coding
   1. Deterministic trends
   2. Stochastic Trends
   3. Unit root tests
   4. Interrupted time series

The fee of $995 includes all course materials.

PayPal and all major credit cards are accepted.

Our Tax ID number is 26-4576270.