3) For a continuous dataset (like the one for Q4 below), can you explain the differences (not to compute the value) between correlation, covariance, and cosine similarity measures of attributes in terms of scale and translation(shift) invariant. ( 10 points) 4) If you decided to do PCA to the following dataset after quality check to reduce the dimension, The following figure shows the variances of each PC. If you need to capture at least 90% of total variance in the data, how many PC5 do you need? After you selected your PCs, what are correlations among the PCs? (10points)