• TUCS Publication Series
• How to Publish in TUCS Series
• Publication Search
• Publication Graph
• Publication Input
• Publication Input Guide
• JuFo Browser

Automated Text Segmentation and Topic Labeling of Clinical Narratives

Hanna Suominen, Sampo Pyysalo, Filip Ginter, Tapio Salakoski, Automated Text Segmentation and Topic Labeling of Clinical Narratives. In: Barbro Back Tapio Salakoski Sanna Salanterä Hanna Suominen Helena Karsten (Ed.), Proceedings of the First Conference on Text and Data Mining of Clinical Documents (Louhi'08), TUCS General Publication, 99-103, 2008.

Abstract:

Electronic patient information systems include numerous functionalities to support clinical judgment and decision-making, but their capabilities to analyze free-text narratives are limited. We apply Hidden Markov Models to divide Finnish intensive care nursing notes into topically coherent segments and assign a topic label to each segment. The method notably outperforms a keyword-based baseline already with a relatively small amount of training data. The result holds the promise of increased information search speed and a more comprehensive overall picture about patients.

Files:

Abstract in PDF-format

BibTeX entry:

@INPROCEEDINGS{inpSuPyGiSa08a,
  title = {Automated Text Segmentation and Topic Labeling of Clinical Narratives},
  booktitle = {Proceedings of the First Conference on Text and Data Mining of Clinical Documents (Louhi'08)},
  author = {Suominen, Hanna and Pyysalo, Sampo and Ginter, Filip and Salakoski, Tapio},
  number = {52},
  series = {TUCS General Publication},
  editor = {Helena Karsten, Barbro Back Tapio Salakoski Sanna Salanterä Hanna Suominen},
  pages = {99-103},
  year = {2008},
  keywords = {text segmentation, topic labeling, Hidden Markov Models, computerized patient records, documentation},
}

Belongs to TUCS Research Unit(s): Turku BioNLP Group

Edit publication