Discussion Of The Application Of Multiple Regression Analysis And Logistic Regression Analysis

In this paper, we mainly discussed the practical application of the multiple regression analysis method belonged to the linear regression method and Logistic regression analysis method belonged to the non-linear regression method. In the first part, we mainly used the influence analysis, the clustering analysis and the multiple regression analysis method and merge these methods together to forecast the runoff of the Wu Qiangxi reservoir. By analyzing the daily precipitation and flow data of 22 hydrology observation stations of Hunan Yuanshui valley including Hexi, Fengtan, Pushi, we first use the statistical method of influence analysis on mass observation data to filtrate, and then employ the clustering method to the mainly influence factor founded after removing the abnormal spot. At last, the multiple regression analysis method is adopted on sorted data to build mathematical model to forecast the runoff of the Wu Qiangxi reservoir. The accuracy ratio of the three models reaches up to 90% and gets a preferably predicting result. In the second part, we apply the Logistic regression analysis way to conduct the research combined with teachers' comments mutually, higher authority leaders'(or experts') appraises to the teacher, students' appraises to the teacher. We propose an overall appraisal model to evaluate the teacher's teaching quality by using the SAS software. This model can appraise the teacher scientifically and be used widely.