Skip to main content


Figure 10 | BMC Bioinformatics

Figure 10

From: Analysis of nanopore detector measurements using Machine-Learning methods, with application to single-molecule kinetic analysis

Figure 10

Classification improvement with Adaboost taking the best 50 from the Inverted-emission 150 feature set. 95% accuracy is possible for discriminating 9GC from 9TA hairpins with no data dropped with use of Adaboost, without Adaboosting, the accuracy is approx. 91%. This demonstrates a significant robustness to what the SVM can "learn" in the presence of noise (some of the 2600 component have richer information, but even more are noise contributors). This also validates the effectiveness with which the 150-parameter compression was able to describe the two-state dominant blockade data found for the nine base-pair hairpin and other types of "toggler" blockades, as well as the utility of the inverted features.

Back to article page