기계학습을 이용한 한국 프로야구 관중 수 예측 모델 개발 (The Development prediction model of Korea Professional Baseball league spectator using machine learning)

서지정보

· 발행기관 : 한국체육과학회
· 수록지 정보 : 한국체육과학회지 / 32권 / 5호 / 547 ~ 558페이지
· 저자명 : 조정환, 석부길

초록

[Purpose] The purpose of this study is to identify the main factors related to the prediction of the number of spectators in Korean professional baseball by using machine learning. [Methods] For the purpose of the study, the daily numbers of spectators for professional baseball from 2017 to 2019 were collected. External factors such as the weather and holidays on the day of the match and the internal situation of the match, such as the away team factor, were input as observation variables. The collected data was analyzed with Python ver 3.6, and the predictive power was cross-validated using three machine learning models: Lasso regression, random forest, and XGboost. [Results] As a result of the analysis, the XGboost model showed the highest predictive power and showed 58.4% accuracy when predicting the number of spectators for the entire KBO league. The most frequently used factor in the entire league was the ‘Date’ factor, and as a single-factor, holidays were the most frequently used in prediction. As for the factors for predicting the total number of spectators by team, the ‘Away team’ factor and the ‘Date’ factor were most frequently used. [Conclusions] Based on the results of this study, it is decided that teams and league will be able to suggest various marketing strategies if the number of spectators is predicted considering the game performance, opponent team, and weather.

영어초록

[Purpose] The purpose of this study is to identify the main factors related to the prediction of the number of spectators in Korean professional baseball by using machine learning. [Methods] For the purpose of the study, the daily numbers of spectators for professional baseball from 2017 to 2019 were collected. External factors such as the weather and holidays on the day of the match and the internal situation of the match, such as the away team factor, were input as observation variables. The collected data was analyzed with Python ver 3.6, and the predictive power was cross-validated using three machine learning models: Lasso regression, random forest, and XGboost. [Results] As a result of the analysis, the XGboost model showed the highest predictive power and showed 58.4% accuracy when predicting the number of spectators for the entire KBO league. The most frequently used factor in the entire league was the ‘Date’ factor, and as a single-factor, holidays were the most frequently used in prediction. As for the factors for predicting the total number of spectators by team, the ‘Away team’ factor and the ‘Date’ factor were most frequently used. [Conclusions] Based on the results of this study, it is decided that teams and league will be able to suggest various marketing strategies if the number of spectators is predicted considering the game performance, opponent team, and weather.

참고자료

· 없음