Using Bayesian network to predict win or lose of the NBA games

• Main Authors: Hau-kai Chuang, 莊皓凱
• Other Authors: 吳榮彬
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/38846931858694297830

碩士 === 逢甲大學 === 統計學系統計與精算碩士班 === 101 === Sports statistics can be said to be more popular, many foreign sports have introduced a data analyst, just to enhance the team, the players and the opponent&;#39;s ability to self-understanding the pros and cons, and using the detailed records of the court to find an important factor to enhance the winning percentage. As can be seen in the sports arena there are still many unknown messages, information, waiting for us to discover discussed. National Basketball Association can be said that now one of the most popular professional sports in the past, we tend to focus on the field than some obvious data, such as free throw attempts, assists. However, these data can not be explained completely utilize their interpretation of a game and the players ability, therefore, offense and defense conversion efficiency value, four factors and other advanced data will be excavated to create eleven, among Dean Oliver made the most famous of the four factors, the use of the court&;#39;s four most influential factors to create the four advanced variables allows us to better understand the rapidly changing pitch, and methods of data analysis are very diverse in recent years, the emergence of Bayesian network analysis of data and forecasts the addition of a choice, the relationships between variables to use to learn and predict, so this research will take this method to NBA2012-13 season, regular season analysis and forecasting, and I would like to discuss the pros and cons of this method the degree.