The Study of Predicting Winning Model for CPBL Teams
| Autor: | Wen,Yang,Deng, 鄧文揚 |
|---|---|
| Rok vydání: | 2013 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 101 The main purpose of this study is to find out which key factors affecting the outcome of the game, and use the key ones to construct a predictive model for the Chinese Professional Baseball League winning baseball games. The data comes from the competitions from Lanew Bears VS Brother Elephants team in 2009. As for the research processes, first the stepwise regression analysis is used to filter out the impact factors of the outcome between Lanew Bears and Brother Elephants teams. The important factors are RBI, innings pitched, the pitcher hits, errors, stolen base, wild pitch, and the pitcher strikeouts. After that, the five prediction models were set up for the study, which are logistic regression prediction model, neural network prediction model, decision tree prediction model, and neural network prediction model with decision tree analysis. The neural network prediction model is a black box system, which is unable to understand the inner cells processing system. Thus, the decision tree method is adopted to have a clearable interpretation. Finally, the results of prediction precision, explainable and reasonable process among the five prediction models are all compared with. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|