Neural Networks Predicting the Winner of a Board Game
BGG Link:
https://boardgamegeek.com/boardgame/173346/7-wonders-duel
A classification task focused on predicting the winner in the two-player board game, 7 Wonders Duel, using game logs from BoardGameArena.com. Two distinct neural network methodologies were developed. An RNN utilized individual moves as input, predicting outcomes based on sequences within the same match. In contrast, a CNN employed filtered and tokenized game logs, leveraging word embeddings for predictions. Parameter tuning was performed on a training set of 1,011 matches, resulting in the best model achieving over 67% accuracy when trained on a total of 2,168 matches, including an untested final set of 434 matches. This approach has implications for educational game design by reducing early-game futility and offering a more efficient alternative to traditional pedagogical methods.
*************************************************************************************
7 Wonders Duel đş https://tinyurl.com/krbahcrw
Abyss đ§ https://tinyurl.com/4fvpdzhp
Blood Rage đż https://tinyurl.com/2p82ky2z
Carcassonne đ° https://tinyurl.com/32wnxbz8
Feast For Odin đš https://tinyurl.com/2p9dty8v
Jaipur đŞ https://tinyurl.com/ycku2sve
Lewis and Clark đś https://tinyurl.com/yckscxvz
Love Letter đ https://tinyurl.com/4umbw8ma
Potion Explosion đś https://tinyurl.com/yeyktfpw
Space Base đ https://tinyurl.com/3r88sbrd
Splendor đ https://tinyurl.com/27266pz9
Stone Age đĽ https://tinyurl.com/yckkudpp
Takenoko đź https://tinyurl.com/2p88ym8b
Welcome To đ https://tinyurl.com/yc2x3zzh
Zooloretto đŚ https://tinyurl.com/3v6vus23
*********************************************************************************