Reputation: 1601
Serie of values might, or might not have clusters. In example:
[1, 2, 1, 0, 1, 1, 0, -1, 0, 21, 22]
has [21, 22] falling into a second cluster.
What unsupervised method allows for telling if there is more than 1 cluster in data without manually setting any threshold?
Upvotes: 0
Views: 32
Reputation: 4983
clusters are somewhat abstract definition, getting it's meaning from the data and context, in your sequence there can be 1 or multiple clusters, depends on what algo you used and how many clusters did you ask it to find
you can use methods to evaluate the output clustering and by this deduce the optimal number of clusters most methods works by compare the similarity between points in the same cluster and in different clusters
I suggest you read the following https://en.wikipedia.org/wiki/Elbow_method_(clustering) https://en.wikipedia.org/wiki/Silhouette_(clustering)
Upvotes: 1