Card Price Prediction of Trading Cards Using Machine Learning Methods

摘要

In this paper, we try to predict the card prices of the trading card game using their information. The trading card game market is growing by the increasing popularity of the board game or the digital card game in the e-sports in recent years. The trading card game is a kind of card game which two or more people plays a card with some text or symbols those characteristics expresses a ruling or interaction to the other card. This interaction of cards may work effectively in the game, prices of those card pairs will be increased with the popularity of its combination. We have a hypothesis that card text is useful for prediction of card prices from the importance of card combinations. Therefore, in this research, we focus on not only the basic card information but also card text. Moreover, we use several machine learning method for prediction of card prices, and we analyze which machine learning method is an effect.