Mining Events through Activity Title Extraction and Venue Coupling

Yuan Hao Lin, Chia Hui Chang, Hsiu Min Chuang

研究成果: 書貢獻/報告類型會議論文篇章同行評審

摘要

In this paper, we discuss the challenges to construct an event/activity search engine through activity title extraction and venue recognition as well as relation coupling. While distant supervision is a common technique to speed up training data preparation, it does not always work on social network. For activity titles, they may contain other entities such as person, venues and temporal expressions, which could be much longer than general named entities (such as person names). Venue recognition is another challenge, as they could be an address, a specific point-of-interest like an restaurant or organization name, or across a neighborhood. Another problem is how to determine the venue and correct date of the event when multiple venues or temporal expressions are recognized in a message. In this paper, a sequential pattern mining approach is applied to discover rules for coupling the recognized place names with the activity title in discussion. The experimental result shows that our approach is beneficial and practical for locating social activity venues.

原文???core.languages.en_GB???
主出版物標題Proceedings - 25th International Conference on Technologies and Applications of Artificial Intelligence, TAAI 2020
發行者Institute of Electrical and Electronics Engineers Inc.
頁面136-141
頁數6
ISBN(電子)9781665403801
DOIs
出版狀態已出版 - 12月 2020
事件25th International Conference on Technologies and Applications of Artificial Intelligence, TAAI 2020 - Taipei, Taiwan
持續時間: 3 12月 20205 12月 2020

出版系列

名字Proceedings - 25th International Conference on Technologies and Applications of Artificial Intelligence, TAAI 2020

???event.eventtypes.event.conference???

???event.eventtypes.event.conference???25th International Conference on Technologies and Applications of Artificial Intelligence, TAAI 2020
國家/地區Taiwan
城市Taipei
期間3/12/205/12/20

指紋

深入研究「Mining Events through Activity Title Extraction and Venue Coupling」主題。共同形成了獨特的指紋。

引用此