Reddit: A Gold Mine for Personality Prediction

Matej Gjurković, Jan Šnajder


Abstract
Automated personality prediction from social media is gaining increasing attention in natural language processing and social sciences communities. However, due to high labeling costs and privacy issues, the few publicly available datasets are of limited size and low topic diversity. We address this problem by introducing a large-scale dataset derived from Reddit, a source so far overlooked for personality prediction. The dataset is labeled with Myers-Briggs Type Indicators (MBTI) and comes with a rich set of features for more than 9k users. We carry out a preliminary feature analysis, revealing marked differences between the MBTI dimensions and poles. Furthermore, we use the dataset to train and evaluate benchmark personality prediction models, achieving macro F1-scores between 67% and 82% on the individual dimensions and 82% accuracy for exact or one-off accurate type prediction. These results are encouraging and comparable with the reliability of standardized tests.
Anthology ID:
W18-1112
Volume:
Proceedings of the Second Workshop on Computational Modeling of People’s Opinions, Personality, and Emotions in Social Media
Month:
June
Year:
2018
Address:
New Orleans, Louisiana, USA
Editors:
Malvina Nissim, Viviana Patti, Barbara Plank, Claudia Wagner
Venue:
PEOPLES
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
87–97
Language:
URL:
https://aclanthology.org/W18-1112
DOI:
10.18653/v1/W18-1112
Bibkey:
Cite (ACL):
Matej Gjurković and Jan Šnajder. 2018. Reddit: A Gold Mine for Personality Prediction. In Proceedings of the Second Workshop on Computational Modeling of People’s Opinions, Personality, and Emotions in Social Media, pages 87–97, New Orleans, Louisiana, USA. Association for Computational Linguistics.
Cite (Informal):
Reddit: A Gold Mine for Personality Prediction (Gjurković & Šnajder, PEOPLES 2018)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-1/W18-1112.pdf