Fig. 3
From: Sketching and sampling approaches for fast and accurate long read classification
Key results across parameters on the CGR dataset. a Classification accuracy across a range of read lengths and error rates, averaged over three simulated runs with variance between runs noted on each bar, b the effect on accuracy of increasing the target number of shared hashes, c the impact read length and error have on sketch size across our approaches, and d the accuracy and overhead of a MinHash approach across a range of k values