This content has been marked as final. Show 4 replies
OCluster only accepts numerical data whereas KMeans will handle both.
Can you check to see if that is the source of the difference?
My table (ID, V1, V2, V3, V4, V5) used to be defined as (Varchar2, float, float, float, float, float). Now, I changed it to (Number, float, float, float, float, float).
Same problem in a different way. Only 700 records out of ~4K were assigned to one cluster!
What do you think?
Hi,1 person found this helpful
Check out Marcos's reply in the following posting.
It explains that this condition can happen given the nature of the data and the algorithm.
O Cluster Algorithm giving only one cluster
Thanks Mark, I guess this explains what is happening here as my values look like below, I guess OC is not able to detect any split in the data because of the small difference between the values. Not sure if this makes any sense!
(ID, V1, V2, V3, V4)
(123, 0.00124, 0.00114, 0.00110, 0.00105, 0.00100)