RATE: Overcoming Noise and Sparsity of Textual Features in Real-Time Location Estimation
This work addresses location inference for applications like localized search and event detection, but it is incremental as it builds on existing topic modeling and feature integration approaches.
The paper tackled the problems of noise and sparsity in textual features for real-time location estimation of social media users, resulting in the RATE algorithm that outperforms benchmarks in region classification precision and mean distance error.
Real-time location inference of social media users is the fundamental of some spatial applications such as localized search and event detection. While tweet text is the most commonly used feature in location estimation, most of the prior works suffer from either the noise or the sparsity of textual features. In this paper, we aim to tackle these two problems. We use topic modeling as a building block to characterize the geographic topic variation and lexical variation so that "one-hot" encoding vectors will no longer be directly used. We also incorporate other features which can be extracted through the Twitter streaming API to overcome the noise problem. Experimental results show that our RATE algorithm outperforms several benchmark methods, both in the precision of region classification and the mean distance error of latitude and longitude regression.