The variable selection process was initiated in Chapter 13, where the univariable procedure was first executed. The univariable procedure was executed in a manner that reduced the number of variables based on statistical significance (p-value <0.25), while also eliminating variables based on data quality (null proportion, zero proportion, and total number of distinct values for nominal variables). In the subsequent Chapter 14, the process of variable selection was advanced by the elimination of redundant variables. These were deemed to be variables that did not significantly contribute to the total variability already explained by their competing variables in the model. This was achieved through the application of the iterative collinearity reduction method, or ICRM. In Chapter 15, we addressed the typical issues of scale, units, linearity, null values, and the inclusion of nominal variables that are inherent in the development of any model. We then demonstrated how these issues can be overcome through the implementation of the WoE transformation. Furthermore, it was demonstrated that the WoE can be employed as an univariable selection method. However, it was advised that this approach should be avoided, as it may potentially disrupt crucial partial associations during the development of the multivariable model. The multivariable selection process is now ready to be initiated. However, it is first necessary to ascertain the number of variables that are anticipated to be included in this phase.