Håvard Kvamme: Gender prediction on Norwegian Twitter accounts

Title: Gender prediction on Norwegian Twitter accounts

Abstract: In this presentation I will give an overview of my master’s thesis, investigating approaches to predicting the gender of Norwegian Twitter accounts. Through Twitter’s public APIs, various account information are available, such as tweets (text), personal descriptions, friends networks and profile images. These unstructured features were processed in different ways, such that individual classifiers could be fitted and then later combined in an attempt to increase the prediction accuracy.