I want to cluster health care profiles based on SVD-based LSI method.Is SVR (singular value rescaling) efficient one or not.Please kindly clarify this.
This link can be helpful: https://scholar.google.com/scholar?q=Latent+Semantic+Indexing+of+SVD-based&hl=en&as_sdt=0&as_vis=1&oi=scholart&sa=X&ved=0ahUKEwjHyuCxvM3aAhUBKMAKHQb8DTcQgQMIBzAA
LSI method is based on SVD and used for matrix rank reduction. However we can use this for dimensionality reduction. In this case, the dimension is always over several thousands.