An active learning approach for clustering single-cell RNA-seq data
python Run_AL.py --data ./Datasets/10X_PBMC_select_2100_top2000.h5 --sn 50 --k 20 --budget 800 --split 0.7 --model SVM --method E --seed 1026
--data: the scRNA-seq data count matrix.
--sn: size of the initial training set.
--k: added cells in each active learning iteration.
--budget: total cells that can be labeled by oracle.
--split: train/test split ratio.
--model: classifer used in active learning model. Options: 1) SVM; 2) LR (logistic regression); 3) RF (random forest); 4) MLP (multilayer proception).
--method: sammple seletion algorithm. Options: E: entropy; M: margin; L: likelihood.
--seed: randomness
Lin, X., Liu, H., Wei, Z., Roy, S. B., & Gao, N. (2022). An active learning approach for clustering single-cell RNA-seq data. Laboratory Investigation, 102(3), 227-235.