The method of establishing process performance model--the card square test _cmmi5 of classification variables

Method of establishing process performance model

Chi-Square test is a widely used hypothesis testing method, often used in the statistical inference of classification data, including: two rate or two composition ratio of the card-square test, a plurality of ratios or multiple composition ratios of the card test and the correlation analysis of the classification data.

The chi-square test can test the consistency between the actual observation times and the theoretical times between the categories of nominal variables of a single polynomial, and the number of observations is the actual count based on the sample data, and the theoretical times are the expected times according to the theory or experience. This kind of test is called the fitting test.

When the correlation judgment between the fixed class data is made, the analysis method of the card square is used to analyze the correlation, and the sampling requirements are at least 5 times.

Its basic data sometimes requires rank conversion and can be manipulated in the rank window of the Minitab.

Figure 1 Row rank operation

After eliminating the multiple collinearity judgment, we can perform the operation of chi-square analysis in Minitab: "Cross grouping table and card side", then the regression model can be obtained.

Fig. 2 Card Square Analysis