SPSS多元回归分析实例(国外英语资料).doc

SPSS多元回归分析实例 Multiple regression analysis In most practical problems, the factors that affect the dependent variable are not one but multiple, and we call this type of regression multiple regression analysis. You can set up the multivariate linear regression model between the dependent variable y and the respective variable xj (j = 1, 2, 3,... n). Among them: b0 is the regression constant; Bk (k = 1, 2, 3,... n) is the regression parameter; E is a random error. Multiple regression in the case of disease prediction: The following four predictors were selected by the correlation coefficient method of the disease station in one area. For the maximum 10 days, the maximum number of moths is a maximum of 10 days. X2 is the total amount of ovulate (lumps) in the middle of April and mid-season. X3 is precipitation (mm) in mid-april, x4 is in mid-april rain day (day); The amount of y (head/m2) is predicted for a generation of stickworm larvae. Table 2-1. The forecast quantity is y: the larva of each square metre is 0 ~ 10 head is 1, 11 ~ 20 head is 2 grade, 21 ~ 40 head is 3, 40 head above is 4 level. Forecast factor: x1 induces the moth from 0 ~ 300 to l, 301 ~ 600 heads 2, 601 ~ 1000 heads 3, 1000 heads to 4 levels; The number of x2 eggs is 1, 15l ~ 300 is 2, 301 ~ 550 is 3, 550 is 4. X3 precipitation 0 ~ 10.0 mm is 1, 10.1 ~ 13.2 mm is 2, 13.3 ~ 17.0 mm is 3, 17.0 mm above is 4. X4 rain day 0 ~ 2 days is 1, 3 ~ 4 days is 2, 5 days is 3, 6 days or 6 days above is 4. Table 2-1 The x1 x2 The x3 x4 y years Amount of moth level Amount of eggs level precipitation level The rain, level The larvae density level 1960 1022 4 112 one 4.3 one 2 one 10 one 1961 300 one 440 3 0.1 one one one 4 one 1962 699 3 67 one 7.5 one one one 9 one 1963 1876 4 675 4 17.1 4 7 4 55 4 1965 43 one 80 one 1.9 one 2 one one one 1966 422 2 20 one 0 one 0 one 3 one 1967 806 3 510 3 11.8 2 3 2 28 3 1976 115 one 240 2 0.6 one 2 one 7 one 1971 718 3 1460 4 18.4 4 4 2 45 4 1972 803 3 630 4 13.4 3 3 2 26 3 197