The purpose of this assignment is to demonstrate steps performed in a K-Means Cluster analysis.

•  Week 4 Cluster Data.xlsx (8.977 KB)

Included with this assignment is an Excel spreadsheet that contains data with two dimension values.

The purpose of this assignment is to demonstrate steps performed in a K-Means Cluster analysis.

Review the “k-MEANS CLUSTERING ALGORITHM” section in Chapter 4 of the Sharda et. al. textbook for additional background.

Use Excel to perform the following data analysis.

1. Plot the data on a scatter plot.
2. Determine the ideal number of clusters.
3. Choose random center points (centroids) for each cluster. (Note:  Each student will select a different random set of centroids.)
4. Using a standard distance formula measure the distance from each data point to each center point.
5. Assign each data point to an initial cluster region based on closeness.
6. For each cluster calculate new center points.
7. Repeat steps 4 through 6.

You will use Excel to help with calculations, but only standard functions should be used (i.e. don’t use a plug-in to perform the analysis for you.)  You need to show your work doing this analysis the long way.  If you were to repeat steps 4 through 6, what will likely happen with the cluster centroids?  

Here is a link to an example spreadsheet using a smaller data set.  It contains two tabs.  The first tab is the raw data.  The second tab contains the analysis that was performed.  Make sure that you use a different starting center points from the example.

Example Excel Analysis