I have a data frame of 500,000x23 dimensions. The data is binary, representing presence or absence. The data follows identified trees through time (23 years) and looks at whether the tree is present that year, or is not. What I am trying to do is cluster trees together that may be behaving similarly to identify spatially, regions showcasing similar patterns.
Initially, I tried A K-means cluster, however, when comparing wss of kmean clusters they would never converge. Upon reading it seems kmeans is inappropriate for binary type data like mine? I have tried to use hierarchical type approach but the dataset is too large, and I run into memory limitation errors.
Are there any other options for clustering data in this type and size, or is there another method to try?
