Skip to main content
. 2022 Nov 30;23:514. doi: 10.1186/s12859-022-05065-3

Fig. 3.

Fig. 3

Performance comparison of the three strategies for the subsampled BRCA datasets. Peak memory, minimum runtime and classification accuracy were measured. Differences in accuracy resulted from the algorithmic implementation specificities of Dask and SPE. The data sizes shown in MB correspond to uncompressed tabular files