Part III. Clustering countries by cultural dimensions

In this part we are going to discover groups of countries based on Geert Hofstede cultural dimensions.

The input data is in Countries.csv. The description of the attributes is in CulturalDimensions.html.

Perform normalization of all numeric values. Remove attribute country name.

Perform K-means (K=2) clustering of countries by cultural dimensions.

Save the output of clustering as Results2.arff. Copy data rows into data file clusterResults2.txt.

In order to properly see 2 clusters of countries, use the code in ClusterAssignments.java. Add this file to a new java project, and put Countries.csv and clusterResults2.txt into the project directory.

Part of Assignment 3. Run K-means algorithm on this dataset with K=3, and K=4. Generate the list of countries in each cluster for each clustering. Which value of K (2, 3, or 4) seems to produce the most meaningful clusters?