An analysis on the applicability of supervised learning algorithms on card games

Abstract

Data mining is the process of extracting patterns from large data sets by combining methods from statistics and artificial intelligence with database management. One of the classes of data mining is classification, where the goal is to generalize a known structure to apply to new data. Supervised learning, a branch of the classification algorithms uses a set of training data to produce an inferred function, called a classifier, which is then used to predict the correct output value for any valid input object. This approach allows supervised learning algorithms to be applied with notable success on games, by predicting the outcome of future matches based on the stored results of previously played matches. We will present the outcome of applying two supervised learning algorithms on a very popular card game, called “Dominion”.