This document discusses machine learning projects for classifying and clustering glass data using R. It first describes classifying glass data using ridge regression and plotting the results. It shows the classification performance is good based on error rate but poor based on ROC. Using higher order polynomials improves ROC, TPR and FPR. It also notes how to properly implement ridge regression. The document then demonstrates clustering glass data using multi-dimensional scaling, k-means clustering, and hierarchical clustering and compares the clustering to the original labels.