Descriptive Data Mining

What types of business insights do you think descriptive data mining could uncover?

How might reducing dimensions help analyze very wide datasets with many variables?

Why is it useful for principal components to be uncorrelated with each other?

What are some potential business applications of clustering you can think of?

How might the choice of distance measure impact clustering results?

What are potential advantages and disadvantages of needing to pre-specify k?

How does hierarchical clustering differ from k-means in its approach?

Which method do you think might be most robust to outliers, and why?

Why is it important to consider multiple criteria when evaluating clusters?

Can you think of an example association rule for grocery store purchases?