Identifying Emotion Labels from Psychiatric Social Texts Using Independent Component Analysis

Proceedings of COLING 2014, the 25th International Conference on Computational Linguistics: Technical Papers, pages 837–847, Dublin, Ireland, August 23-29 2014. Identifying Emotion Labels from Psychiatric Social Texts Using Independent Component Analysis Liang-Chih Yu1,2 and Chun-Yuan Ho1 1Department of Information Management 2Innovation Center for Big Data and Digital Convergence Yuan Ze University, Chung-Li, Taiwan lcyu@saturn.yzu.edu.tw, s986304@mail.yzu.edu.tw  Abstract Accessing the web has been an efficient and effective means to acquire self-help knowledge when suf- fering from depressive problems. Many mental health websites have developed community-based ser- vices such as web forums and blogs for Internet users to share their depressive problems with other us- ers and health professionals. Other users or health professionals can then make recommendations in re- sponse to these problems. Such communications produce a large number of documents called psychiat- ric social texts containing rich emotion labels representing different depressive problems. Automatically identify such emotion labels can make online psychiatric services more effective. This study proposes a framework combining latent semantic analysis (LSA) and independent component analysis (ICA) to ex- tract concept-level features for emotion label identification. LSA is used to discover latent concepts that do not frequently occur in psychiatric social texts, and ICA is used to extract independent components by minimizing the term dependence among the concepts. By combining LSA and ICA, more useful la- tent concepts can be discovered for different emotion labels, and the dependence between them can also be minimized. The discriminant power of classifiers can thus be improved by training them on the inde- pendent components with minimized term overlap. Experimental results show that the use of concept- level features yielded better performance than the use of word-level features. Additionally, combining LSA and ICA improved the performance of using each LSA and ICA alone. 1 Introduction Sentiment analysis has been successfully applied for many applications (Picard, 1997; Pang and Lee, 2008; Calvo and D'Mello, 2010; Liu, 2012; Johansson and Moschitti, 2013; Balahur et al., 2014). Analysis of online psychiatric or mental health texts (Wu et al., 2005; Yu et al., 2009) is also an emerging field that could benefit from sentiment analysis techniques because more and more people search for help from the web when they suffered from depressive problems, which boost the develop- ment of online community-based services for Internet users to share their depressive problems with other users and health professionals. Through these services, individuals can describe their depressive symptoms via web forums and blogs. Other users or health professionals can then make recommenda- tions in response to these problems. Figure 1 shows an example psychiatric social text collected from PsychPark (http://www.psychpark.org), a virtual psychiatric clinic, maintained by a group of volunteer professionals belonging to the Taiwan Association of Mental Health Informatics (Bai et al., 2001; Lin et al., 2003). This example shows a subject’s depressive problems and the responses recommended by the experts. Some meaningful tags called emotion labels herein are also annotated by the experts to indicate which categories the text belongs to. These emotion labels are useful …

Content truncated to 500 words