Exploration of Clusters
First, we shall add a few components to the workflow from the previous assignment
- Attach Select Columns to Hierachical clustering.
- Attach Naive Bayesian classifier to Select Columns.
- Attach Nomogram to Naive Bayesian classifier.
Now for some work. We shall first explore properties of one group vs. the rest of the world.
- Double click the connection between Hierarchical Clustering to Select Columns. Connect Selected Data from clustering to Data in Select Columns.
- In Select columns, designate Selected as target. With these two steps, you the naive Bayesian classifier will construct a model for distinguishing between selected and unselected countries.
- In hierarchical clustering, select the cluster with developed countries (the one including the eastern and western Europe and others).
- Observe the nomogram. (Set Target class to "Yes".) What is typical for these countries?
- Repeat this exercise for the cluster with African countries.
Next, let us compare different clusters.
- Double click the connection between Hierarchical Clustering to Select Columns. Connect Data from clustering to Data in Select Columns.
- In Select columns, designate Cluster as target. With these two steps, you the naive Bayesian classifier will construct a model for distinguishing between the selected clusters.
- In hierarchical clustering, select two clusters: the predominantly western and predominantly eastern Europen cluster.
- In the nomogram, select one cluster or the other and observe the differences.
Last modified: Monday, 14 March 2022, 11:10 AM