The Web Science Trust

Self-Supervised Learning for Medical Web Disease Reporting Events Detection

Stewart, Avaré and Nejdl, Wolfgang (2011) Self-Supervised Learning for Medical Web Disease Reporting Events Detection. pp. 1-2. In: Proceedings of the ACM WebSci'11, June 14-17 2011, Koblenz, Germany.

[img]
Preview
PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
151Kb
[img]
Preview
Image (PNG)
32Kb

Official URL: http://www.websci11.org/fileadmin/websci/Posters/196_paper.pdf

Abstract

In Web Science, Social Media Based Epidemic Intelligence has emerged as a type of medical intelligence gathering that supports health ocials in routinely identifying potential health threats from more dynamic sources of information. State-of-the-art supervised approaches for SM-EI suer from the high costs associated with manually labeling training examples. This paper addresses the aforementioned problem by build- ing a self-supervised classier, one that labels its own train- ing examples. Our results show that a self-supervised classi- er, which discriminately selects its support vectors at each iteration achieves an F-measure of 78%. These results are comparable with existing, state-of-the-art systems, which rely exclusively on labeled data to build a classier.

Item Type:Conference or Workshop Item (Poster)
Web Science Comments:WebSci Conference 2011
Subjects:WS1 Computer Science
Web Science Events > Web Science 2011
ID Code:462
Deposited By: Lisa Sugiura
Deposited On:07 Jun 2011 16:43
Last Modified:25 Oct 2011 17:12

Repository Staff Only: item control page

EPrints Logo
Web Science Repository is powered by EPrints 3 which is developed by the School of Electronics and Computer Science at the University of Southampton. More information and software credits.