Skip to main content
Fig. 6 | BMC Bioinformatics

Fig. 6

From: Keeping up with the genomes: efficient learning of our increasing knowledge of the tree of life

Fig. 6

NBC Incremental learning classifier (NBC++) training/update time is evaluated by the same simulated dataset. This figure shows the update time for incremental learning method (in blue) and traditional full update (NBC) method (in black). The different components of the algorithm are shown – the time that the core NBC algorithm takes (solid) vs. that time plus the average time per genome for Jellyfish to count k-mers (dashed). We can see that the k-mer counting alone can start to add a significant amount of computation when recomputing the entire training dataset

Back to article page