Contributed to the sklearn library framework for building a class of multi-modality analysis functions. I solely implemented OCCA module for orthogonal canonical correlation analysis which is a variant of standard CCA(Canonical Correlation Analysis) family of multi-modla analysis algorithms which is a linear multivariate statistical analysis method for dimensionality reduction and feature fusion for better analysis and evaluation of multivariate data. It tries to find to mutually independent feature set for 2 different modalities for feature extraction and fusion by preserving the orthogonality constraint even for large number of reduced dimensions.
The entire package along with a thourough analysis repo has been solely developed and contributed by me on a public master repo of CCA consisting of various other variants of CCA along with OCCA, which could be found at: https://github.com/shekhar-sharma/DataScience/tree/main/Groups/Group_ID_14