A User Study on Public Health Events Detected within the Medical Ecosystem

authored by

Avaré Stewart, Eelco Herder, Matthew Smith, Wolfgang Nejdl

Abstract

The great influx of Medical-Web data makes the task of computer-assisted gathering and interpretation of Social Media-based Epidemic Intelligence (SM-EI) a very challenging one. State-of-the-art approaches usually use supervised machine learning algorithms to gather data from a variety of sources in this medical ecosystem, mining this data for specific event patterns and information discovery. Supervised approaches not only limit the type of detectable events, but also requires learning examples be given to the machine learning algorithm in advance. On the other hand, the more generic and flexible unsupervised machine learning methods currently produce such complex results, that the domain experts are not capable of assessing the results in a natural and efficient manner. In this paper, we present a novel framework with which SM-EI field practitioners can interact with medical ecosystem data, and assess the results of such complex unsupervised SM-EI algorithms. The assessment framework and the unsupervised epidemic event detection algorithm have been fully implemented and a quantitative study is presented to show the validity of the new approach to SM-EI.

Organisation(s)

L3S Research Centre

Type

Conference contribution

Pages

127-132

No. of pages

6

Publication date

02.11.2011

Publication status

Published

Peer reviewed

Yes

ASJC Scopus subject areas

Computer Graphics and Computer-Aided Design, Computer Networks and Communications, Environmental Engineering

Sustainable Development Goals

SDG 3 - Good Health and Well-being

Electronic version(s)

https://doi.org/10.1109/DEST.2011.5936610 (Access: Unknown)