MLB 프로젝트 - 케이스스터디

MLB 경기예측 정확도가 60%만 넘어가도 굉장히 인상적인 수치다

The best existing scientific literature (Elfrink, Valero, Jia, etc.) has largely produced accuracy levels within the 57–59.5% range. At face value this may seem unexceptional; however, a barometer often considered the best predictive metric for professional sports — the Vegas betting odds — achieved an accuracy level of merely 58.2% over the past six years (calculated using data from covers.com). Given the high degree of randomness inherent to our problem, predicting nearly 60% of games correctly is actually quite impressive

결국은 홈팀이 이길 것이냐 질것이냐를 판단하는 문제이다

predicting the outcome of a baseball game is a binary classification problem, namely, whether or not the home team will win or lose

VEGAS BET을 기준으로 저자는 정확도를 비교하였다

=> 데이터 뿐만 아니라 유명 베팅사이트의 지난 BET을 크롤링해오면 이 또한 하나의 지표가 될거라 판단됨

https://www.oddsportal.com/baseball/usa/mlb/results/

MLB Results & Historical Odds, Baseball USA Archive

 

 

출처 : https://garretnourse.medium.com/a-machine-learning-algorithm-for-predicting-outcomes-of-mlb-games-fa17710f3c04

A Machine Learning Algorithm for Predicting Outcomes of MLB Games

https://github.com/GarretNourse/Machine-Learning-Predict-MLB-Games

GitHub - GarretNourse/Machine-Learning-Predict-MLB-Games

https://fisher.wharton.upenn.edu/wp-content/uploads/2020/09/Thesis_Andrew-Cui.pdf

https://github.com/kylejohnson363/Predicting-MLB-Games-with-Machine-Learning

GitHub - kylejohnson363/Predicting-MLB-Games-with-Machine-Learning: Combine classification algorithms to predict the winner of e