Zhang, Thomas, Jason HD Cho, and Chengxiang Zhai. "Understanding user intents in online health forums." Proceedings of the 5th ACM Conference on Bioinformatics, Computational Biology, and Health Informatics. 2014.

This paper addresses the challenge of extracting useful information from the large, unstructured content on online health forums by understanding user intents. A taxonomy of intents is developed to capture user information needs, and novel pattern-based features are proposed for use with a multiclass support vector machine (SVM) classifier to classify thread posts according to their underlying intents. A dataset of 1,200 Health-Boards posts from four forum topics was manually labeled by three annotators for this task. Experimental results demonstrate that the SVM with pattern-based features effectively identifies user intents, achieving a maximum precision of 75%. The classifier's performance remains comparable even when training and testing on posts from different forum topics. Additionally, the trained classifier was applied to a MedHelp dataset to analyze the distribution of intents across various forum topics.