부산대학교 도서관

The Korean Ladies Professional Golf Association (KLPGA) is one of the most influential women's professional golf tours in the world, and the qualification to play until the end of the tournament, determined by whether a player passes the cut or not, is important for both players and industry. Recently, with the development of artificial intelligence and big data, research on predicting sports game results has been actively conducted and industrially utilized. However, research on predicting cut off in KLPGA tournaments is still scarce. This study aimed to address this research gap by applying deep learning algorithms to predict players' cut off in KLPGA tournaments. To achieve this, data was collected from KLPGA website and built prediction models based on three deep learning algorithms (LSTM, CNN, mWDN) to empirically predict players' cut outcomes. The empirical results showed that the multi-level wavelet decomposition network (mWDN) was the most effective deep learning algorithm for predicting KLPGA players' cut outcomes, with an F1-score of 0.75. When comparing the predicted player list based on deep learning algorithms with the actual list of players who passed the cut in the top three tournaments with the highest prize money held during the 2023 KLPGA tour, it accurately predicted the cut outcomes of 276 out of 352 players (78.4%). This study holds academic significance and practical implications for the golf industry as the first in South Korea to empirically predict KLPGA players' cut outcomes using deep learning algorithms and derive a practical list of players.