Machine Learning vs. Statistical Methods A Key Focus

M4, the fourth of the well-known Makridakis Competitions that have run since the 1980s and believed to be the world’s first forecasting competitions, is still open for submissions from any team globally (academic, corporate or individual). There are currently over 200 teams from 42 countries participating, with the majority of the teams hailing from the USA, India, the UK and Brazil. This latest edition of the M-Competition will be the largest of its kind and remains open to all avid forecasters. The deadline for submissions is 31 May 2018.

Professor Spyros Makridakis, a Professor at the University of Nicosia and one of the pioneers in the field of forecasting, is leading the design and assessment of the competition. M4 is organized by the Institute For the Future (IFF) at the University of Nicosia (UNIC), with the support of the Forecasting & Strategy Unit at the National Technical University of Athens (NTUA).

M-Competitions more generally seek to identify the most accurate forecasting method(s) for different types of predictions. Participants in M4 specifically need to make forecasts across 100,000 real-life time-series.  Of particular interest in M4 is whether Artificial Intelligence (AI) / machine learning methods will perform as well as traditional statistical forecasting methods.

M4 is offering over €25,000 in cash prizes to the winners, as well as honorary prizes for the method that achieves the most accurate confidence intervals forecasts.   The prizes are generously provided by UNIC, Professor Makridakis personally, Uber and Amazon.

Specifically, the prizes are:

  • UNIC Prize for Best Performing Method | €10,000 | University of Nicosia
  • Makridakis Prize for Best Performing Method | €10,000 | Professor Spyros Makridakis
  • UBER Student Prize for Best Performing Method by (a) Student(s) | €5,000 | UBER
  • Amazon Prize for Best Reproducible Forecasting Method | €2,000 | Amazon

Complete information about the M4 can be found on the website. Additional information can also be found in the user guide.