I want to get overlapped clusters and I knew that fuzzy c-means can give me the required results.
But when I use it it gives me disjoint clusters
I have changed the no. of clusters and the fuzzifier no. It also gives me disjoint clusters
Ian working on string data (on apply clustering on documents)
What should i do to get the overlapped clusters ?
Attached my workflow
please check it and tell me what to do
thanks in advance
The clusters are disjoint, but a point does not have a fixed assignment to each cluster.
Each point has the probability to contain to a cluster and the Winner cluster is the one where the data point is the nearest two.
Basically every data point contains to every cluster.
Thanks so much Iris for your comment
I want the winner cluster to be more than one cluster, In other words, the document logically can be assigned to more than one cluster example: one document can be categorized to the Information retrieval cluster and at the same time is categorized to Machine learning cluster (so, the winner cluster is IR & Machine learning)
I want to get this result tell me how please
Is there a clustering node in KNIME that can make overlapping clusters?
The second question is about the steps mentioned in my workflow for document clustering, are my steps correct or Is there any feedback?