Question on K-pod algorithm with only one point in one cluster

[lizhenstat]

Thanks for you great working on implementation of k-POD, which provides us an alternative way for clustering on data points with missing values.

However, I have one related question on this method:
I use k-POD as a method for my problem of clustering data.
Dataset (for example): number of data points=200, number of features=10, number of cluster=3, level of missing value is around 20%
And I notice that k-POD method gives me three cluster assignment(for example)

data-1 --> cluster-1 (1 point)
data-2 --> cluster-2 (1 point)
data-3, data-4,..., data-100 --> cluster-3 (98 point)

My problem is that I want to at least two points in one cluster, do you have some suggesstion to solve this problem?

Any help would be appreciated and thanks for your time !