Multivariate Analysis - Clustering - Multivariate Analysis Assignment Help

Internal Code : 1AJBJF

Multivariate Analysis Assignment Help

TASK Cluster units into groups using hierarchical clustering. Compare at least 3 different combinations of methods and dissimilarity measures, one of which must be Ward’s method with squared Euclidian distance. For all methods compute the value of Ward’s criterion function at the selected number of clusters. Select the most appropriate method (not (only) based on criterion function, which is in some sense “subjective”). Cluster units into groups using k-means algorithm. Use both results of hierarchical clustering and “scree plot” based on criterion function at different number of clusters. Use the procedure where the starting leaders/centers are selected at random Compare partitions to each other based on criterion function and by assessing similarity of partitions (using contingency table and (Adjusted) Rand Index, at least one pair). Select the best partition and use this partition in the remainder of the assignment. Select the best partition and interpret it (group sizes and their meaning – based on averages of original and standardized variables.)