/Prediction-of-LoL-Tier-with-Summoner-Name

๐Ÿ‘พ ๋ฆฌ๊ทธ์˜ค๋ธŒ๋ ˆ์ „๋“œ ์†Œํ™˜์‚ฌ ์ด๋ฆ„ ๊ธฐ๋ฐ˜ ํ‹ฐ์–ด ์˜ˆ์ธก & ์›Œ๋“œํด๋ผ์šฐ๋“œ ๋ถ„์„

Primary LanguageJupyter Notebook

Prediction-of-LoL-Tier-with-Summoner-Name

RNN(Recurrent Neural Network)์„ ์ด์šฉํ•œ (์žฌ๋ฏธ๋กœ ๋ณด๋Š”) ๋ฆฌ๊ทธ์˜ค๋ธŒ๋ ˆ์ „๋“œ ์†Œํ™˜์‚ฌ ์ด๋ฆ„์œผ๋กœ ํ‹ฐ์–ด ์˜ˆ์ธกํ•˜๊ธฐ

(+๋ถ€๋ก : ๋ฆฌ๊ทธ์˜ค๋ธŒ๋ ˆ์ „๋“œ ์†Œํ™˜์‚ฌ ์ด๋ฆ„ ๊ธฐ๋ฐ˜ ์›Œ๋“œํด๋ผ์šฐ๋“œ ๋ถ„์„)

>>> tier_predict('KT way')
('๋งˆ์Šคํ„ฐ ์ด์ƒ', 0.25377932)

>>> tier_predict('ํ”Œ๋ž˜ํ‹ฐ๋„˜ ๋ฌธ์ง€๊ธฐ')
('ํ”Œ๋ž˜ํ‹ฐ๋„˜', 0.21506898)

>>> tier_predict('์•„๋ชฌ๋“œ๊ฐ€ ์ฃฝ์œผ๋ฉด')
('๋‹ค์ด์•„๋ชฌ๋“œ', 0.28879693)

>>> tier_predict('์•„์ด์–ธ๋งจ')
('์•„์ด์–ธ', 0.18356498)

>>> tier_predict('์ •๋ณดํ†ต์‹ ๊ธฐ์ˆ ํ˜‘ํšŒ')
('ํ”Œ๋ž˜ํ‹ฐ๋„˜', 0.18466833)

์‚ฌ์šฉํ•˜๊ธฐ

๋ชจ๋ธ ํ›ˆ๋ จ

๋ชจ๋ธ ์‚ฌ์šฉ

์‚ฌ์šฉ ๋ฐ์ดํ„ฐ

  • ์ „์ฒด 3,718,157๊ฑด
    • ๋ฆฌ๊ทธ์˜ค๋ธŒ๋ ˆ์ „๋“œ ์†Œํ™˜์‚ฌ ๋žญํ‚น
    • ๋ฆฌ๊ทธ์˜ค๋ธŒ๋ ˆ์ „๋“œ ์†Œํ™˜์‚ฌ ์ด๋ฆ„
    • ๋ฆฌ๊ทธ์˜ค๋ธŒ๋ ˆ์ „๋“œ ํ‹ฐ์–ด
ranking nickname tier
1 JUGKlNG challenger
2 viper3 challenger
3 KT Way challenger
4 SMB heixiaohu challenger
5 ๋ƒฅ๋˜ฅ๋ฒŒ๋ ˆ challenger
... ... ...

ํ‹ฐ์–ด ๋ณ„ ๋ถ„ํฌ

  • Challenger tier : 296๊ฑด
  • Grandmaster tier : 681๊ฑด
  • Master tier : 3,806๊ฑด
  • Diamond tier : 57,195๊ฑด
  • Platinum tier : 404,496๊ฑด
  • Gold tier : 1,020,424๊ฑด
  • Silver tier : 1,272,286๊ฑด
  • Bronze tier : 854,653๊ฑด
  • Iron tier : 104,320๊ฑด

์‚ฌ์šฉํ•œ RNN ๋ชจ๋ธ ๋ณ„ ์„ฑ๋Šฅ

๊ฐ ๋ชจ๋ธ ๋ณ„ ๋’ค์— ๋ถ™์€ ์ˆซ์ž๋Š” embedding_dim, hidden_units, batch_size ๊ฐ’์„ ์˜๋ฏธํ•จ.

RNN Models Accuracy
LSTM-128 0.21945
LSTM-64 0.21972
BiLSTM-128 0.22147
BiLSTM-64 0.22132
GRU-128 0.22105
GRU-64 0.21993
BiGRU-128 0.22192
BiGRU-64 0.22210

2022๋…„ ์ธํŠธ์•„์ด ์—ฌ๋ฆ„ ํ•ด์ปคํ†ค "๊ทธ๋กœ์šฐ์—…ํ†ค" - ์ธ๊ณต์ง€๋Šฅ ํŒ€ ์ฐธ๊ฐ€ ์ž‘ํ’ˆ

Hits