James Landrith says it's very unclear:
While the stats most often quoted show extremely low numbers of female predation, the reality differs. Quite often the same act committed by a female as by a male is counted separately or not included in official tabulations at all depending on the statistical model. These models, with all of their obvious built-in bias, are then parroted around as if they are apples to apples comparisons of male and female predation. As such biases and outright distortions are often used to eliminate them from from data sets or intentionally isolate such data in lesser or hidden categories, we have no real idea of just how many female predators exist today.
The Dish featured Landrith's account of his own rape here.
The Dish 