Year: 2015
Type: dissertation
Abstract: In this thesis, methods for predicting the gender of Norwegian Twitter\naccounts were investigated. Through Twitterâ s public APIs, various\naccount information is available. Tweets (text), personal descriptions,\nfriends networks, and profile images were the main fields investigated.\nFirst separate classifiers were fitted to features from the different fields,\nand later the individual classifiersâ posterior probability estimates were\ncombined to achieve increased accuracy. The datasets were labeled\nthough comparison of the accountsâ names and names in the Norwegian\npopulation. Subsets of accounts with very gender specific names were\nused for training and testing. The highest balanced accuracy obtained\nwas around 0.89. This, however, required access to the accountsâ profile\nimages (85% of the data). Without images, the accuracy dropped to\naround 0.85.
Source: BIBSYS Brage (BIBSYS (Norway))
Author: Håvard Kvamme
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
SDG: Gender equality
Open Access status: green
OpenAlex ID: W2466230984