1. Determine the variables and draw a scatter chart
Intuitively determine which mathematical regression model is selected from the scatter chart
2. select an appropriate mathematical regression model to establish a sample equation.
The mona1 equation is usually used: Y = a + bt
3. Use the least square method to calculate coefficients A and B
Select data> data analysis> regression in Excel. Of course, you need to load the macro first, otherwise you will not see the data analysis menu.
4. Determine the sample Regression Equation
5. Inspection
1) t-test
If the p-value is less than the obvious level (default: 5%), it indicates that there is a linear relationship between the variable X and Y. The coefficient A and B are verified by the obvious level, and the coefficients A and B are significant.
2) F test
If significance F is less than the obvious level (default: 5%), it indicates that the regression equation passes the general significance test, that is, the X and Y correlation, and the overall regression equation is significant.
3) R2 Test
The closer R square is to 1, the better the regression equation fit.
6. Prediction
Prediction Using regression equations that pass tests