Analyze Complete Data
Now that your imputed values appear to be satisfactory, you are ready to run an analysis on the "complete" data. The dataset contains a variable Customer category [custcat] that segments the customer base by service usage patterns, categorizing the customers into four groups. If you can fit a model using demographic information to predict group membership, you can customize offers for individual prospective customers.
- Activate the telcoImputed dataset. To create a multinomial
logistic regression model for the complete data, from the menus choose:
Figure 1. Multinomial Logistic Regression dialog - Select Customer category as the dependent variable.
- Select Marital status, Level of education, Retired, and Gender as factors.
- Select Age in Years, Years at current address, Years with current employer, Number of people in household, and Log of income as covariates.
- You want to compare other customers to those who subscribe to the
Basic service, so select Customer category and click Reference category.
Figure 2. Reference Category dialog box - Select First category.
- Click Continue.
- Click Model in the Multinomial Logistic Regression
dialog box.
Figure 3. Model dialog box - Select Custom/Stepwise.
- Select Main effects from the Stepwise Terms Build Term(s) dropdown.
- Select lninc through reside as Stepwise Terms.
- Click Continue.
- Click OK in the Multinomial Logistic Regression dialog box.