Extracting Issue-Based Information System Structures from Online Discussions for Automated Facilitation Agent
รหัสดีโอไอ
Creator Atsuya Sakai
Title Extracting Issue-Based Information System Structures from Online Discussions for Automated Facilitation Agent
Contributor Shota Suzuki, Rafik Hadf, Takayuki Ito
Publisher Sirindhorn International Institute of Technology, Bangkadi Campus (SIIT-BKD)
Publication Year 2565
Journal Title Journal of Intelligent Informatics and Smart Technology 
Journal Vol. 7
Page no. 1-7
Keyword Natural Language Processing, Machine Learning, Argument Mining, Deep Learning, Social Networks, Conversational Agents, Automated Facilitation
URL Website https://ph05.tci-thaijo.org/index.php/JIIST
Website title Journal of Intelligent Informatics and Smart Technology 
ISSN 2586-9167
Abstract Automated facilitation agents are currently being used to enhance the quality of online discussions. To this end, agents need to extract meaningful text units before applying facilitation rules. The issue-based information system (IBIS) is a viable way to extract and classify text from online discussions. In this paper, we propose a novel approach for text classification by adopting the IBIS system for online discussions. The method starts by identifying the correct IBIS structures and then finds the correct type of nodes in the structures. The approach relies on a Graph Attention Network (GAT) for both tasks in order to directly learn the IBIS structure. That is, the GAT encodes the graph structures and then classifies different structures using an attention architecture. To evaluate the performance of the approach, we conducted a set of experiments on a persuasive essays dataset formatted with the IBIS model. The experimental results show that the proposed approach is able to accurately classify the structures and the text nodes in online discussions.
Sirindhorn International Institute of Technology, Bangkadi Campus

บรรณานุกรม

EndNote

APA

Chicago

MLA

ดิจิตอลไฟล์

Digital File
DOI Smart-Search
สวัสดีค่ะ ยินดีให้บริการสอบถาม และสืบค้นข้อมูลตัวระบุวัตถุดิจิทัล (ดีโอไอ) สำนักการวิจัยแห่งชาติ (วช.) ค่ะ