Sistema automatizado de tomada de decisão em um jogo de poker

Abstract

Poker is a sport with huge growth and prominence worldwide. In this context, thousands of athletes already practice poker professionally while several technologies related to the area are being developed. Also, in evidence in the twenty-first century, data mining and machine learning emerge as excellent mechanisms for performing tasks that only a few years ago would be performed by humans, giving machines that power. In this work, a decision-making mechanism will be developed in a poker game with the use of algorithms linked to the subarea of artificial intelligence, machine learning, for data mining tasks. Its implementation includes several concepts of programming and statistics, combined to achieve the highest expected value in a No Limit Holde'm - 6 max poker game, focusing on pre-flop decisions. During the work, the best actions for each position of the table (from a total of six positions) will be defined, considering the possible previous actions of the opponents, quantity of chips and the cards received by the player.