The report suggests that globally no correlation lies between the
datasets. Thus, if there is no correlation, the datasets are random at
best globally. But, from Kernel Densities we see that certain local
clusters were being formed. It is thus important to check whether there
are outlier cluster being formed in these data for us to allow that
there is some correlations in certain parts. The cluster analysis thus
allows us to look for outlier (local I’s) in this data. This thus allows
for looking at certain sets of clusters within the data.
After the cluster analysis proves that there are certain outliers in the
data, the Low High outliers provide certain information to us as they
overlap in regions of high Kernel Density as well.
The following are the statistics of the cluster analysis for kernel
densities: