Running the analysis (linear models)
- To build a linear model, from the menus choose:
Figure 1. Linear models fields tab 
This dataset is also used to build a model to predict fraudulent claims, so Cost of claim in thousands (claim_amount) is defined in the dataset as an input; so when you open the linear models dialog, it is initially placed in the Predictors (Inputs) list. Note that the Target list is empty because Fraudulent claim (fraudulent) has a nominal measurement level, and linear models are only valid for continuous targets.
- Remove Cost of claim in thousands (claim_amount) from the predictors
list and select it as the target.
For the remaining fields in the Predictors list, categorical (nominal, and ordinal) fields are used as factors in the model and continuous fields are used as covariates.
- Click Run.