Skip to main content
. 2013 Sep;195(1):181–193. doi: 10.1534/genetics.113.152587

Figure 4.

Figure 4

Feature weights learned by parameter-specific SVMs. (A) Feature weights of the SVM trained on (s = 0.04, τ = 700), a regime in which Tajima’s D is sensitive. (B) Feature weights of the SVM trained on (s = 0.04, τ = 2000), a regime in which Fay and Wu’s H is sensitive. We note that the rightmost feature (representing fixed differences) has no equivalent in D or H.