2.2 Gender Identification Approaches
In our experiment, we consider sixteen approaches to gen-
der identication. All gender identication approaches can
report female", male" or unknown"4.
The rst group of approaches is used as the baseline for
comparison with the more advanced techniques. The sim-
plest approach is the straw man that always predicts male".
Next we consider two collections of name-based heuristics:
genderComputer [17]5; and the most popular gender identi-
cation tool on Mashape Gender Guesser.6
To address RQ1, i.e., name-based gender identication in
absence of rst name frequency lists, we propose to exploit
a (general-purpose) social network as a source of name fre-
quency information. Indeed, if by consulting such a social
network we can establish that most individuals called Orit"
are female, then we can reasonably conclude thatOrit"from