Papers › Neural Demographic Prediction in Social Media with Deep Multi-View Multi-Task Learning
Neural Demographic Prediction in Social Media with Deep Multi-View Multi-Task Learning
Yantong Lai, Yijun Su, Cong Xue, Daren Zha
Utilizing the demographic information of social media users is very essential for personalized online services. However, it is difficult to collect such information in most realistic scenarios. Luckily, the reviews posted by users can provide rich clues for inferring their demographics, since users with different demographics such as gender and age usually have differences in their contents and expressing styles. In this paper, we propose a neural approach for demographic prediction based on user reviews. The core of our approach is a deep multi-view multi-task learning model. Our model first learns context representations from reviews using a context encoder, which takes semantics and syntactics into consideration. Meanwhile, we learn sentiment and topic representations from selected sentiment and topic words using a word encoder separately, which consists of a convolutional neural network to capture the local contexts of reviews in word-level. Then, we learn a unified user representation from context, sentiment and topic representations and apply multi-task learning for inferring user’s gender and age simultaneously. Experimental results on three real-world datasets validate the effectiveness of our approach. To facilitate future research, we release the codes and datasets at https://github.com/icmpnorequest/DASFAA2021_DMVMT.
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
Results from the paper archive 2025-07-28
No leaderboard rows for this paper in the archive.
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections