To Analysis of the Zulu Principle and ARIMA Model

• Main Authors: YE SHU-MEI, 葉淑媚
• Other Authors: Sheu Tian-Wei
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/49351079724991325535

碩士 === 國立臺中教育大學 === 教育測驗統計研究所 === 97 === In this research, we use method of the Zulu Principle of Jim Slater to select the company. The data of this research come from the Over-the-Counter during January 2003 to December 2008, and use the ARIMA to find the best predicted model of this company. Moreover, this method may also apply in mathematic exams scores to find the best predicted model. And using the 95 percent confidence interval and mean absolute percent error to measure the effect upon predicted of the model, which one is the best predicted model. The empirical result shows that： （1）The best predicted model of St.Shine is . It predicted value all fall in 95% confidence interval, and mean absolute percent error is 0.8696% （2）The best predicted model of mathematics examination scores is . And, the predicted value all fall in 95% confidence interval, and mean absolute percent error is 3.7%. The St.Shine is compared with mathematics examination scores, and find its error is smaller than mathematics examination scores. But both of them are all fall in 95% confidence interval. So, the predicted model of St.Shine is better than mathematics examination scores.