[2K6-GS-2] Machine learning

[2K6-GS-2-03] Trading Card Price Prediction using Deep Learning with Multimodal Data

〇Kazuki Yabuuchi¹, Naoki Mori², Makoto Okada² (1. Osaka Prefecture University, 2. Osaka Metropolitan University)

Keywords:Deep Learning, Machine Learning, Time Series Analysis, Multimodal Data, Imbalanced Data

Recently, the prediction of change in time-series data by artificial intelligence has attracted much attention. The prices of cards used in trading card games (TCG) are one of the time-series data, such as stock prices and exchange rates. TCG is commonly played between two players and in which players use special cards with power and toughness, rules text that explains special effects. It is popular because of not only the game itself but also the characteristic of the price change. Various factors cause price change, including the data related to game results and rarity, release time, and past prices. From the above background, some studies have been conducted to predict TCG card price change using machine learning. However, few studies are using the deep learning method. In addition, the TCG cards whose prices increase are the minority group. Therefore, approaches to imbalanced data are necessary for training. We confirmed the effectiveness of the proposed methods that predict each card price increase using deep learning with multimodal data.