search

KrishnaswamyLab / MAGIC

MAGIC (Markov Affinity-based Graph Imputation of Cells), is a method for imputing missing values restoring structure of large biological datasets.
GNU General Public License v2.0
341 stars 97 forks source link

RMagic alpha parameter #140

Closed daliabarkley closed 5 years ago

daliabarkley commented 5 years ago

Hello, I apologize for this very basic question. The paper and tutorials use t, k and ka as parameters, whereas the R function uses t, k and alpha. Is alpha simply another name for ka, or is it something different? How would you recommend finding an optimal value for alpha? Thank you very much!

dvdijk commented 5 years ago

we introduced alpha in the new version of magic. alpha determines the decay of the kernel with a larger alpha having a shrper decay. as alpha goes to infinity the kernel becomes unweighted or just a knn kernel. we dont use ka anymore. k and t are the same you can either leave alpha as default or you can increase or decrease depending on if we want a more localized (bigger alpha) or wider (smaller alpha) kernel

daliabarkley commented 5 years ago

Thank you very much for your fast and clear response!