Predicting Results for the Basketball Game based on Data Mining Techniques

• Main Authors: Yi Chen Liu, 劉奕辰
• Other Authors: Min-Sheng Lin
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/gf8nzd

碩士 === 國立臺北科技大學 === 電機工程系所 === 104 === The National Basketball Association (NBA) is one of the most successful professional sports organizations and the highest level of basketball organization in the world. The results of each regular season game may be related to whether the team can qualify for the playoffs. It is one of the key issues for each professional team to dig valuable information from thousands of game statistics and apply to decision-making effectively. This thesis applies data mining techniques to investigate the correlation between the basketball game statistics and basketball game results and try to find the relation between the accuracy of predictions and the time when predictions occur in the regular season. The experimental results show that predicting the results for top teams and bottom teams are more likely accurate than predicting the results for average teams. Besides, predicting the results for bottom teams are more and more accurate as the regular season goes on, but the reverse trend occurs in the cases of top teams.