Hashtag Recommendation with Attention-Based Neural Image Hashtagging Network

Gaosheng Wu, Yuhua Li, Wenjin Yan, Ruixuan Li, Xiwu Gu, Qi Yang

Abstract: With the increasing number of microblog users, the hashtag recommendation task has become an important component in social media. Most hashtag recommendation related methods get relative low precisions, because hashtags are not necessarily related to the content of tweets, which makes hashtag recommendation more challenging. In this work, we propose a new sequence-to-sequence method named attention based neural image hashtagging network (A-NIH) to model sequence relationship between social images and hashtags. To the best of our knowledge, this is the first work that applies attention mechanism to the image-only hashtag recommendation tasks. Our experimental results on the real-world social image dataset shows that our model performs better than the state-of-the-art methods.