Figure 2From: Gendered behavior as a disadvantage in open source software developmentComparing gender inferring algorithms. Accuracy of our gender inference against a baseline and two alternative methods. Precision (a) measures for each category how many categorized items are relevant, and recall (b) captures how many relevant items are selected from all good ones, F score (c) takes the harmonic average of precision and recall, reaches 1 when both metrics are perfectBack to article page