Fundamentals of Statistics contains material of various lectures and courses of H. Lohninger on statistics, data analysis and here for more.

Exercise - Classification of unknown wine samples by PCA

Suppose you are an importer of Italian wine who has received a new supply of wine. Opening and tasting the wine makes you sceptical. The label of the wine does not correspond to the contents of the bottle. So you go to your lab (fortunately you are an analytical chemist, too) and analyze two samples of two different bottles. Here are the results of your analysis:

                 Sample 1       Sample 2
alcohol          13.48          14.06
malic-acid       1.67           2.15
ash              2.64           2.61
alcali-ash       22.5           17.6
magnesium        89             121
phenol-total     2.6            2.6
flavanoids       1.1            2.51      0.52           0.31
proanthcya       2.29           1.25
color intensity  11.75          5.05
hue              0.57           1.06
OD280OD315       1.78           3.58
proline          620            1295

Now take the data set WINE and go to the  DataLab  in order to compare the unknown samples with the analysis of known wines. Try to use principal component analysis in order to get an idea of the unknowns.

Last Update: 2012-10-08