Random Lasso & Hi-Lasso Packages | R & Python

Professor's page: http://ksuweb.kennesaw.edu/~mkang9/

Regression analysis algorithms are becoming increasingly more fine-tuned for inferring gene expression in bioinformatics. Ridge regression was a robust starting ground for this series of regression analysis algorithms, but these have since improved in both accuracy and speed. Some examples of regression analysis that give a chronological scope of the progression this field are: lasso, net elastic, adaptive lasso, random lasso, and hi-lasso.

There is a lack of resources both educational and algorithmic on Random Lasso. Often researches will not attempt to implement Random Lasso due to its complexity, and opt for a less accurate but simple algorithm. Our goal is to release a highly malleable Random Lasso package for both Python and R. As an extension of this, we will also be releasing a package for Doctor Kang's new state-of-the-art Hi-Lasso algorithm. Despite lasso based regression analysis often taking hours to complete, we have ambitions to make these algorithms practical for personal-computers by taking advantage of parallel computing and vectorization.