Figure 6
From: MiningABs: mining associated biomarkers across multi-connected gene expression datasets
Scalability of improved c-LMs. A) Accuracy as a function of various ESCC training datasets, grouped by gene numbers. B) Accuracy as a function of various HCC training datasets, grouped by gene numbers. C) Average running times of panel A and B. Error bars indicate standard error of the means. Improved: improved c-LMs; Defective: removing one out of an improved c-LM-contained genes by turns.