Hashtag2Vec: Learning Hashtag Representation with Relational Hierarchical Embedding Model
/**
- This project is developed by the Intelligent Information Processing Lab, Nankai University, Tianjin, China. (http://dm.nankai.edu.cn/)
- It follows the GPLv3 license. Feel free to use it for non-commercial purpose and please cite our paper:
- @inproceedings{Hashtag2Vec,
- author = {Jie Liu and Zhicheng He and Yalou Huang},
- title = {Hashtag2Vec: Learning Hashtag Representation with Relational Hierarchical Embedding Model},
- booktitle = {Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence, {IJCAI} 2018, July 13-19, 2018, Stockholm, Sweden.},
- pages = {3456--3462},
- year = {2018},
- doi = {10.24963/ijcai.2018/480},
- }
- Contact: jliu@nankai.edu.cn, hezhicheng@mail.nankai.edu.cn */
To run the Hashtag2Vec model, use the codes in the cn.edu.nk.iiplab.hzc.hashtag2vec package. Hashtag2Vector.java : the Hashtag2Vec model with a runnable synthetic example in the main function; Hashtag2Vector_agg.java : the Hashtag2Vec_agg model with a runnable synthetic example in the main function; Logger.java : a tool class to save the running logs; Performance.java: the evaluation metrics for the learned embeddings.
For further project developments, use the tool classes in the cn.edu.nk.iiplab.hzc.basic package. cn.edu.nk.iiplab.hzc.basic.MF : some matrix operation functions implemented in a multi-thread way; cn.edu.nk.iiplab.hzc.basic.thread : thread classes for cn.edu.nk.iiplab.hzc.basic.MF; cn.edu.nk.iiplab.hzc.basic.matrix : dense and sparse matrices; cn.edu.nk.iiplab.hzc.basic.struct : some basic data structures; cn.edu.nk.iiplab.hzc.basic.util : some tool functions.