The data mining models based on the Microsoft Clustering algorithm is targeted towards identifying the relationships between different entities of the dataset and dividing them into logically related groups. This algorithm differs from other algorithms in such a way that these do not require any predictable columns as their prime motive is to identify the groups of data, rather than to predict the value of an attribute. These groupings can then be used to make predictions, identify exceptions, and so on. Thus, the prime usage of this algorithm lies mainly in the data analysis phase where the focus is mainly on the existing/current data to test our hypothesis about the relationships between entities in the data and determine any exceptions (hidden relationships).
The following screenshot shows a data mining model based on the Microsoft Clustering algorithm. This can be seen in the SSDT Mining Models tab.
An important observation regarding the preceding screenshot...