My professor suggested a comparison of various forecasting models as a topic for my semester project. Given that my only experience in statistics is the intro course in probability and statistics taught here. It would really help me out a lot if someone could recommend a book I can follow to read about statistical forecasting and the implementation of such models.

  • 4
    $\begingroup$ Reminder to close-voters: we do have a references tag, and questions about references are on-topic. Also, this is definitely not a case of a self-study question in the sense that closure reason is meant. $\endgroup$ Jan 10, 2022 at 16:34
  • 5
    $\begingroup$ Start with what has already been done: the Hyndman paper referenced at stats.stackexchange.com/questions/559361 is freely available, non-technical, and easy to read. It describes the results of forecasting competitions over the last 40 years. $\endgroup$
    – whuber
    Jan 10, 2022 at 16:35

1 Answer 1


Interestingly enough, we don't seem to have a general thread on references in forecasting (but perusing this search would still be helpful).

There are a number of forecasting textbooks out there. I would order them in three tiers of increasing depth.

First, there are general overviews for people who need a rough understanding of what is happening when we forecast, without necessarily forecasting themselves. The target audience might be managers of forecasters, or MBA students in supply chain optimization or similar fields of study, or people thinking about entering the field of forecasting and wanting to see what would lie ahead of them.

Second, there are resources for the practicing forecaster, who should know how to decide between different forecasting methods, and what the advantages and disadvantages of each method are, but do not necessarily need to understand precisely how the parameters of an ARIMA(p,d,q) model are fitted. This category actually probably covers 80-90% of academic researchers in forecasting.

Third, there are textbooks that will go into the deep details. These typically presuppose understanding of statistics and/or numerical optimization on a graduate level.

The ML community is unfortunately rather unaware of the state of the art in the forecasting field, which has historically been dominated by statisticians and econometricians. Hewamalage et al. (2022, DMKD) summarize best practices in forecast evaluation, aiming specifically at ML experts and data scientists without specific expertise in forecasting.

You might also want to take a look at Foresight: The International Journal of Applied Forecasting, which is a practitioner-oriented forecasting journal. They also publish compilations of past articles on specific topics.

Petropoulos et al. (2021), "Forecasting: Theory and Practice" (recently published in the International Journal of Forecasting) aims at giving a very short overview for many, many aspects of forecasting. It has been placed online as the Forecasting Encyclopedia, and the authors hope to update it continuously. I expect some authors to be more diligent about revisiting this site than others, but it may still stay reasonably well up-to-date.

You might also be interested in the International Symposium on Forecasting, an annual conference on forecasting that draws academics, practitioners, consultants and software vendors, often with practitioner tracks, usually with workshops offered prior to the conference itself. This is a great opportunity to get in contact with forecasters from all walks of life.

Full disclosure: I'm one of the authors of the first books mentioned above, a Deputy Editor for Foresight, and one of the authors of the Petropoulos et al. article.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.