Clustering example
Clustering is one of the Data Mining algorithms implemented in BI2M. The Clustering algorithm query in BI2M should be defined by two steps:
- choosing the case. A case is the basic unit, which will be analyzed by the algorithm
- choosing the characteristics, on which the algorithm will form clusters.
You can start Clustering module using the main menu of BI2M - click File-> New-> Clustering. Choose the desired OLAP cube and the Data Mining wizard appears.
Example:
The database FoodMart 2000 with the OLAP cube Sales is given. We are interested in finding 3 segments of the customers of FoodMart stores in order to create a program for offering different benefits for the customers depending on their personal characteristics. The goal is to increase their loyalty to the stores. We will use the Clustering algorithm in the FoodMart 2000 database that segments the customers in the OLAP cube Sales into three categories based on the following information: Gender, Marital Status, Yearly Income, Education, Member Card, and Store Sales.
Step 1 – As we will group customers, we have to choose Customer as a case on the first page of the OLAP Data Mining Wizard.
Step 2 – at this step we choose the characteristics which will be processed by the algorithm. On their basis the clusters will be created. In current task we are interested in Customers’ Gender, Marital Status, Education, Yearly income Member Card and Stores Sales, that is why we select them.
Set the number of clusters = 3:
The result of data mining is:
