| Summary: | 碩士 === 國立屏東大學 === 應用數學系碩士班 === 107 === The study chose the native pitchers of each team of the Chinese Professional Baseball League (CPBL) from 2014 to 2016 seasons as research objects, exploring the correlation and difference between the performance of the pitchers and the salary. Then, the study used the performance of the pitchers to predict salary and establish the model of the salary. The statistical methods applied in this research were independent-sample t-test, Pearson product-moment correlation analysis and regression analysis. The significant level of the statistical tests in this research were set at α=0.05. The results are concluded as follows:
A. There are three variables in the age, years of CPBL, and years of trade overseas of the pitchers, which are significantly correlated with the performance of the pitchers. There is also a significant difference in the average number of performances of trade overseas pitchers and overseas pitchers without trade. No variable are significantly correlated with background variable or the mean is significantly different in three years.
B. There are three variables in the age, years of CPBL and years of trade overseas of the pitchers, which are significantly positively correlated with the salary of the pitchers. There are two variables in the age and the years of travel, which are significantly positively correlated with the salary in the three years. Then, the average salary of the trade overseas pitchers is better than those without trade overseas experience.
C. Among the 35 pitcher performance variables, there are 19 are significantly correlated with the salary.
D. The performance of pitcher has a significant predictive effect on the salary.
a. Salary model
1. The basic model: establishing the model of the salary is used a stepwise method to select independent variables.
(1) The model in 2015: "SV", "CG", "age", "W" and travel experience were selected. The predictability is 85.0%.
(2) The model in 2016: "year of travel overseas", "age", "travel experience", "WP", "SV", "IP" and "SHO" were selected. The predictability is 86.1%.
(3) The model in 2017: "W", "age", "SV" and "TP" were selected. The predictability is 77.0%.
2. The smallest model: the model consists of "SV" and "age". The predictability from 2015 to 2017 is 61.9%, 33.3% and 56.1% respectively.
3. The specific model: the model consists of "SV", "W" and "age". The predictability from 2015 to 2017 is 79.0%, 63.0% and 69.8% respectively.
b. The salary model for the year can be used to predict the salary for the next year, and the specific model is the best.
c. The important variables which influence salary are "SV", "age" and "W".
|
|---|
