Dear all,
I'm doing a cluster analysis of binary variables in STATA.
Next I have to search for the most suitable number of cluters.
For example here I try to use 5 clusters:
I would like to use the elbow method to optimize the number of clusters.
The question is: Is there a way to compute the explained variability and then graph it for the different numbers of clusters?
Thank you all in advance.
Gianfranco
I'm doing a cluster analysis of binary variables in STATA.
Code:
cluster averagelinkage var1 var2 var3 var4 var5 var6 var7 var8 var9 var10 var11, measure(Gower2)
For example here I try to use 5 clusters:
Code:
cluster gen ak=gr(5)
The question is: Is there a way to compute the explained variability and then graph it for the different numbers of clusters?
Thank you all in advance.
Gianfranco