Start Date
11-8-2016
Description
The Internet has tremendously facilitated the exchange of health care information among users for their different personal health care needs and well-being issues. User generated content in various social media has become an alternative for expressing users’ needs and discovering knowledge. Previous researches mostly focused on finding answers or potential answerers to facilitate the interactions in CQA and most of them worked with open domain. Although researchers have long examined ways to augment questions and answering systems in medical domain, most of them focused on clinicians’ use of the systems rather than general users. This study focuses on questions themselves to investigate the research questions: (1) What are the features and classification techniques useful to predict whether a CQA question can obtain answers from experts? (2) Can the performance of the predict model benefit from the ensemble technique? To what extent can the ensemble strategies improve the system performance?
Recommended Citation
Lin, Yi-Ling; Chung, Cheng-Yu; Kuo, Che-Wei; and Chang, Te-Ming, "Modeling Health Care Q&A Questions with Ensemble Classification Approaches" (2016). AMCIS 2016 Proceedings. 6.
https://aisel.aisnet.org/amcis2016/Health/Presentations/6
Modeling Health Care Q&A Questions with Ensemble Classification Approaches
The Internet has tremendously facilitated the exchange of health care information among users for their different personal health care needs and well-being issues. User generated content in various social media has become an alternative for expressing users’ needs and discovering knowledge. Previous researches mostly focused on finding answers or potential answerers to facilitate the interactions in CQA and most of them worked with open domain. Although researchers have long examined ways to augment questions and answering systems in medical domain, most of them focused on clinicians’ use of the systems rather than general users. This study focuses on questions themselves to investigate the research questions: (1) What are the features and classification techniques useful to predict whether a CQA question can obtain answers from experts? (2) Can the performance of the predict model benefit from the ensemble technique? To what extent can the ensemble strategies improve the system performance?