This project aims to investigate the potential of emotion-based features in the gender identification task, which has been unnoticed by researchers so far. The experimental study is carried out on the texts of two widely used OSNs, Facebook and Twitter.
The experimental study of this paper is centered around three research questions given below.
RQ1: Whether emotions revealed in the Facebook posts differ for men and women.
RQ2: Whether the emotional behavior of men and women exhibited in the OSN content vary across different platforms.
RQ3: Whether the emotion-based attributes could assist in designing a gender predictor classifier
Features:::
Features Description Range f1 to f8 fear, angry, sad, joy, surprise, disgust, trust 0 -1 f9, f10 positive_fraction, negative_fraction 0 -1 f11, f12 #emotion_categories, emotion_variance 0 -8 f13 class_label 0 or 1