Continuing problem 15, the file P17_20.xlsx contains the same wine data. Instead of trying to use a classification algorithm to classify wines into the three known types (A, B, and C), it is interesting to see if a clustering algorithm can discover these known categories. The file contains the results of two runs of a Microsoft Data Mining Detect Categories algorithm (not discussed here). Of course, neither uses the Type variable in column A. The first run didn't specify the number of categories, and the add-in found 7, with category membership in column O of the Data sheet. The second run specified the number of categories as 3, with category membership in column P of the Data sheet. Analyze the results closely. Do either (or both) of the runs seem to discover the known A, B, and C types?