AI predicts team to win the 2018 World Cup

The international team uses AI programs to run 100,000 simulations and calculate the Spanish team with the highest chance of championship.

An artificial intelligence program predicts the 2018 World Cup champion based on 100,000 simulations, according to Science Alert. Using the database from previous tournaments and three different techniques to handle the numbers, the international team of researchers identified the Spanish team to win the final. The results of the study are published June 13 in arXiv.org magazine.

Although most of today's predictions are inclined to the German team, the AI ​​program analyzes both strength and journey to the finals of the teams. Although Germany is capable of beating Spain, the model that indicates the German team will surely encounter many opponents that are harder to play during the tournament."Through the analysis of the winning rate in each round of the tournament, Spain has a slightly better advantage than Germany, mainly because of its ability to stop in Germany's 16th round relatively high , " said the team. rescue said.

The probability of Spain's 2018 World Cup championship is slightly higher than that of Germany according to AI calculation results.

The research team consists of scholars from Dortmund Technology Geology, Munich Technological University in Germany, and Ghent University in Belgium. They use three techniques for data processing, including Poisson regression (interpretation and evaluation of various variables), ranking methods (assessing the current strength of each team) and random forest methods .

For the World Cup, random forest methods are particularly suitable. It can be structured as a football tournament, using machine learning to map random branches for decision trees, corresponding to different scenarios and accompanying consequences. By programming for AI to focus on certain factors when making decision trees, random forest methods can help statisticians better understand the most important factors for the end result, such as for The players that Germany and Spain will surely meet.

The team ran 100,000 simulations for 2018 World Cup. Spain ranked highest with 17.8% chance of victory, followed by Germany (17.1%), Brazil (12.3%), France (11.2%) and Belgium (10.4%) ). Meanwhile, Saudi Arabia is expected to have no chance of reaching the finals.

Goldman Sachs financial company also ran AI program to predict World Cup results and said that the Spanish team will win the championship. Goldman Sachs's calculations also use random forest methods.