Skip to main content
. 2019 Apr 25;2019:baz045. doi: 10.1093/database/baz045

Figure 2.

Figure 2

Feature Binning. To bin a set of terms, we first partition the continuous probability interval [0,1] into equal-spaced sub-intervals, each of width w. Two terms Inline graphic and Inline graphic are grouped into one bin if and only if the probability Inline graphic falls into the same probability sub-interval as Inline graphic, and Inline graphicfalls into the same probability sub-interval as Inline graphic.