Detect negative event for depression tendency from web blogs

Chia Ming Tung, Wen Hsiang Lu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Many negative events happened in short period of time would trigger depressive disorder. We categorized these negative events into four groups, which are Family, Study, Jobs, and Affection. In this paper, we try to identify negative event in the web blog posts through proposed Enhanced Event Extraction method that includes Enhanced Lexicon Feature, POS pattern, and event-emotion pair. Term frequency and length had been used to calculate importance score for a long negative term in Enhanced Lexicon Feature. In POS pattern, we combined multiple POS term as negative event pattern, such VC+Nh, VC+Na, P+VC. We also consider about the distance between event terms and emotion terms. If emotion terms could be labeled in a post, we could use distance relationship to find event around emotion term. Experimental results show that the accuracy rate of Enhanced Event Extraction method is 54.9%.

Original languageEnglish
Title of host publicationThe 15th International Conference on Biomedical Engineering, ICBME 2013
EditorsJames Goh
PublisherSpringer Verlag
Pages801-804
Number of pages4
ISBN (Electronic)9783319029122
DOIs
Publication statusPublished - 2014 Jan 1
Event15th International Conference on Biomedical Engineering, ICBME 2013 - Singapore, Singapore
Duration: 2013 Dec 42013 Dec 7

Publication series

NameIFMBE Proceedings
Volume43
ISSN (Print)1680-0737

Other

Other15th International Conference on Biomedical Engineering, ICBME 2013
CountrySingapore
CitySingapore
Period13-12-0413-12-07

All Science Journal Classification (ASJC) codes

  • Bioengineering
  • Biomedical Engineering

Fingerprint Dive into the research topics of 'Detect negative event for depression tendency from web blogs'. Together they form a unique fingerprint.

Cite this