pyrifo
Reputation: 107
Column- or row-wise hierarchical clustering using Scipy
I'm doing a hierarchical cluster-analysis of a dataframe with python Scipy, and was wondering if the clustering is column-wise-oriented or row-wise-oriented? I get different results depending on whether I transpose the dataframe or not before calculating the distance matrix.
I read somewhere that the "standard" for clustering is to do it column-wise, however I can't seem to find anything about this in the scipy-documentation. Anyone who knows?
Upvotes: 0
Views: 378
Answers (1)
pyrifo
Reputation: 107
For anyone who's wondering the same thing; basic testing with a simple dataframe showed that it is row-wise
Upvotes: 1
Related Questions
- Scipy hierarchical clustering appropriate linkage method
- With SciPy how do I get clustering for k=? with doing hierarchical clustering
- Hierarchical clustering for categorical data in python
- Using scipy hierarchical clustering with objects
- Convert distance pairs to distance matrix to use in hierarchical clustering
- Hierarchical clustering on sparse observation matrix
- Hierarchical clustering a pairwise distance matrix of precomputed distances
- Clustering with Scipy in Python? (hierarchical clustering)
- Hierarchical clustering from confusion matrix with python
- Use Distance Matrix in scipy.cluster.hierarchy.linkage()?