@techreport{oai:ipsj.ixsq.nii.ac.jp:00206878,
 author = {松田, 真治 and 伊東, 栄介 and Shinji, Matsuda and Eisuke, Ito},
 issue = {4},
 month = {Sep},
 note = {本研究では麻雀における他家の待ち牌予測を，機械学習手法の Neural Network を用いて行う．先行研究では，他のプレイヤーの待ち牌を予測する場合，時系列情報を省いている．そのため，手順や局面の状況を利用できない予測となっていた．手順や局面の状況を予測に使うため，本研究では牌の切られた順序情報を学習に用いる手法を提案する．具体的に，麻雀ゲーム「天鳳」の牌譜データを対象に，先行研究の手法と本研究の提案手法とを適用し，待牌を予測した．その結果，先行研究と比較して F1 score を 4.16 % 向上させることができた．, In this study, we apply a machine learning method, Neural Network, to estimate the waiting tiles of other players in Mahjong game. In a previous research, they omitted time-series information when estimating the waiting tiles of other players. As a result, they are unable to use the situation of the procedure or the phase of the game. In this study, we propose a method to use the information on the order in which the tiles are cut in order to make predictions. Specifically, we applied our proposed method and the method of the previous study to the data of the Mahjong game “Tenho” to estimate the waiting tiles. As a result, we were able to improve the F1 score by 4.16 % compared to the previous study.},
 title = {麻雀における他家の待ち牌推測},
 year = {2020}
}