Applied Forecasting Course

Duration Workload Mode of Study Live Q&A
6 weeks 10h/week Online training Wednesdays
at 18:00 (CY time)

Why Register?

About the Course

THE STARTING DATE FOR THIS COURSE IS 18th October 2021 –

Register Now!

The Applied Forecasting Course is a six-week online course, (next intake 18th October 2021) covering all types of forecasting methods including time series and regression. The course offers individuals and businesses concrete insights on how to improve their accuracy realistically and estimate the uncertainty in their forecasts, while considering its implications to risk. In addition to the traditional statistical forecasting methods, Machine Learning ones like Neural Networks and Decision Trees, will also be covered.

The most important advantage of the course is its emphasis on hands-on learning by encouraging participants to use actual data to both predict and estimate the uncertainty of their forecasts and its implication to risk. This course will offer participants the opportunity to harness 40 years of Professor Makridakis’ knowledge and experience, and master forecasting in six weeks. In addition, Dr. Cirillo will cover extreme events and fat-tails, and their implications to forecasting and uncertainty, while Dr. Spiliotis will present both the traditional statistical forecasting methods, and the more advanced, Machine Learning ones.

Course Outline

Getting started with R

To be provided as a recorded lecture before the beginning of the course for anyone lacking a programming background in R. Students will be able to ask questions during the course.

Taught by Pasquale Cirillo

During Week 1 there will be a preparatory session, Session 0, during which, a tutorial will be offered titled Introduction to R. This contains some videos and reading materials meant to help students familiarise themselves, with R, the programming language of the course. In Session 0, you will also find some hints and tips to get the most out of the Applied Forecasting course. During Session 0 there will be no live classes. You will be provided with a recorded lectureA pre-knowledge of R is not necessary.

Students will be able to ask questions during the live sessions.

All course material will be available online one week before the official starting date of the course.

Live classes will start in Week 1 Session1.

Session 0: Recorded Tutorial: Getting started with R

Taught by Pasquale Cirillo

Session 1: Time Series Decomposition

Seasonality, trend, cycle and randomness, data relationships

Taught by Spyros Makridakis

Session 2: Forecasting and Uncertainty

Understanding, measuring and dealing with various types of uncertainty

Taught by Spyros Makridakis

Session 3: The M Competitions

Benchmarks, simple vs. sophisticated methods, combining forecasts, computational costs versus accuracy, the end of forecasting winter, simple ML methods

Taught by Spyros Makridakis

Session 4: Statistical Forecasting Methods

Naïve methods, exponential smoothing models, and the Theta method

Taught by Evangelos Spiliotis

First bi-weekly assignment (to be submitted at the end of week 4)

Session 5: Linear Regression

Using explanatory variables to predict the future

Taught by Spyros Makridakis

Session 6: Machine Learning, Deep Learning, Cross Learning, and Hybrid Models

An introduction to Machine Learning, its variants, and its state-of-the-art implementations

Taught by Evangelos Spiliotis

Session 7: Advanced Machine Learning Methods

Neural Networks and Regression Trees

Taught by Evangelos Spiliotis

Session 8: Case study

Application of Machine Learning methods in energy prices forecasting

Taught by Evangelos Spiliotis

Second bi-weekly assignment (to be submitted at the end of week 6)

Session 9: Extremes and Fat tails

Taught by Pasquale Cirillo

Session 10: Tail risk and modeling

Taught by Pasquale Cirillo

Session 11: Some successful forecasting applications 

Taught by Pasquale Cirillo

Session 12:  The limits of forecasting

Taught by Pasquale Cirillo

Final assignment (to be submitted two weeks after the end of the course

What Will You Learn?

The course will last six weeks, and it will cover the following topics among others:

Where to start and how to apply forecasting in your business.

Estimating the future uncertainty in your predictions and taking concrete actions to deal with such uncertaintyTime Series forecasting and its use, utilizing available, free software programs.

Where to start and how to apply forecasting in your business.

How to estimate the future uncertainty in your predictions and take concrete actions to deal with such uncertainty, while considering its implications to risk.

Time series analysis, decomposition, and judgmental adjustments.

The use of statistical time series forecasting methods, like exponential smoothing and Theta.

Regression models and their planning value.

The basics of Machine Learning, like Neural Networks and Decision Trees, for time series forecasting.

Ways for improving forecasting accuracy through the combination of forecasts.

How to deal with extreme events and fat tails.

How to exploit the findings of the M Competitions to improve the forecasting function of your organization.

How to forecast and estimate uncertainty using the open free statistical language R. (Tutorials for those unfamiliar with R will be provided to help them in successfully completing the assignments of the course).

Learn about key forecasting applications, like risk management, inventory management, sales and operations, budgeting, and long-term growth.

Apply what you have learnt to your own data, to improve your business decisions.

Who Should Take This Course?

The course is designed for individuals working in finance, marketing, sales, retail, and operations, as well as consultants. Key applications covered in this course are from the above areas, using real-life data to illustrate the value added by both standard and state-of-the-art forecasting methods. Participants will acquire a practical understanding of how forecasting can help them to improve the accuracy of their predictions and estimate uncertainty correctly with an emphasis on what needs to be done in practice to reap the maximum benefits from the systematic usage of forecasting techniques.

In terms of prerequisites, a basic understanding of statistics is sufficient. No particular programming skill is conversely required, as all the necessary knowledge will be covered during the course.

All in all, the most important prerequisite is the willingness to learn how to become a good forecaster.

How Will You Be Assessed?

Participants will be asked to work on two short assignments (one to be submitted at the end of the 4th week and the second at the end of the 6th week) and a final assignment (to be submitted two weeks after the end of the course), where participants will be asked to analyze a set of real data, which can come either from their own business environment (preferred, if available) or can be provided by the instructor instead.

All assignments will be reviewed by the teaching faculty. Only those who have successfully submitted the assignments will be eligible for the certificate of completion, which will be digitally verifiable.
The certificate will be issued in your name at no additional cost, upon successful completion of the course.
There is no extra cost for maintaining your certificate.

If, during the course, you decide not to work on the assignments, you can still access all course materials and participate in all activities, but no certificate will be issued.

Register Today