Edit

Context

With the Edit function, you can modify existing Classes to add or remove nodes as their members. This is often very helpful for fine-tuning Classes after having performed Variable Clustering.

Usage, Example & Workflow Illustration

The following example is based on a typical driver analysis study of customer satisfaction (see Seminar: Key Drivers Analysis and Optimization). One key element in this study’s workflow is to group manifest variables into so-called clusters using BayesiaLab’s Variable Clustering function. In BayesiaLab, clusters are implemented as Classes, which means that a “node belonging to a cluster” translates into that node being a member of a Class representing the cluster. In this workflow example, we reassign the node Wiper Performance (front/rear), which was originally assigned to the Class [Factor_9], to the Class [Factor_4]. We perform this reassignment based on our domain knowledge, i.e., “overrule” the automatic assignment produced by Variable Clustering. Although the motivation is irrelevant to our example, we might think this particular node fits better into Class [Factor_4].