A self-guided, weakly supervised learning algorithm for feature extraction from noisy and high-dimensional data. The method facilitates the identification of patterns that reflect underlying group structures across all samples in a dataset. It incorporates a novel strategy to integrate spatial information, enhancing the interpretability of results in spatially resolved data.
This third version of KODAMA will be available soon on https://CRAN.R-project.org/package=KODAMA.
library(devtools)
install_github("tkcaccia/KODAMA")