Three Free Online Non-Forecasting Courses

As mentioned in my recent Foresight article (Energy Forecasting: Past, Present and Future), virtually all types of energy forecasts are connected. I believe that the best way to further advance our knowledge in such an interdisciplinary field is to take an interdisciplinary approach. Statistics is obviously a core subject in energy forecasting. On the other hand, many other important subjects are essential for developing useful forecasts, such as mathematics, electrical engineering and computer science. In this post, I'm going to introduce three courses from MIT OpenCourseWare that are very helpful to energy forecasters. All of these three courses have video lectures available.

Linear Algebra
There is no need to further emphasize the importance of linear algebra in forecasting. When pursuing my PhD at NC State University, I took two matrix and linear algebra courses from Carl Meyer. I couldn't find any other linear algebra course better than Meyer's. His textbook Matrix Analysis and Applied Linear Algebra, is also very well written. This linear algebra course from Gilbert Strang is the best alternative in my opinion. It is also one of the most popular courses in MIT OCW.

Artificial Intelligence
Believe it or not, in the scientific community, most energy forecasting papers talk about artificial intelligence techniques and their combinations and variations. Although the practical value of those papers is very close to zero, artificial intelligence is not of no use in energy forecasting. The key is to understand these techniques well enough to apply them in the right business environment. Patrick Henry Winston not only covers a solid introduction to many AI techniques, he also offers many insights and words of wisdom useful for research in general.

Introduction to Algorithms
Accuracy is not the only measure of forecasting performance. In many situations, it is not even the most important measure. Any approach nice-looking on the paper has to be translated to some computer language when it comes to production. This course introduces the analysis and design of computer algorithms. In today's market, many tools and solutions already take care of data structure in the back end, so that the programming skills seem to be unnecessary to forecasters. Nevertheless, the forecasters who know how the computer would handle the models definitely have a better chance of developing efficient approach than their counterpart.