1. Select at least 3 large (with over 500 instances) regression datasets (the output is numerical instead of binary or discrete class) from the UCI Machine Learning Repository, or datasets from the WEKA website.
You will apply the following WEKA regression algorithms on these datasets.
2. Generate two synthetic 2D datasets (with at least 1,000 data points each) that you think K-means clustering algorithm would work and would not work respectively. Use simpleKMeans in Weka to cluster the two datasets several times and visualize the results to see whether simpleKMeans really works or not. Try different k values. Based on the results and your original hypothesis, please point out whether the results are consistent with your hypothesis, and try to analyze the reason why simpleKMeans fails to work or works well.
Note: To generate the synthetic datasets, you can use tools such as Matlab or you can write a program in whatever language you prefer.
Please submit files containing the two synthetic datasets, your code that generates the data, visualization of your clustering results in Weka, and the analysis of the results.