The objective of the main-effects model is to integrate the optimal models derived from the given subset of solution spaces into a unified multivariable model, which will serve as the foundation for a subsequent model-building process. In contrast to the preceding model-building processes, which yielded 21 model variants, this new process will commence with the automatic binning process (see Figure 18-3). Thereafter, the flow will be bifurcated into two principal branches. One of these branches will be connected directly to a scorecard node, employing the backward elimination only (BE-Only) multivariable selection method. The second branch will be connected to an iterative collinearity reduction method (ICRM) node, which will in turn be divided into two branches; one branch will utilize the BE-Only method, while the other will employ the all-subsets multivariable selection method. Subsequently, the main-effects building process will yield three variants: