Twitter User Geolocation Using a Unified Text and Network Prediction Model
Computation and Language
2015-09-23 v3 Social and Information Networks
Abstract
We propose a label propagation approach to geolocation prediction based on Modified Adsorption, with two enhancements:(1) the removal of "celebrity" nodes to increase location homophily and boost tractability, and (2) he incorporation of text-based geolocation priors for test users. Experiments over three Twitter benchmark datasets achieve state-of-the-art results, and demonstrate the effectiveness of the enhancements.
Cite
@article{arxiv.1506.08259,
title = {Twitter User Geolocation Using a Unified Text and Network Prediction Model},
author = {Afshin Rahimi and Trevor Cohn and Timothy Baldwin},
journal= {arXiv preprint arXiv:1506.08259},
year = {2015}
}
Comments
To appear in ACL 2015, Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics (ACL 2015)