Mining rare cases in post-operative pain by means of outlier detection

  • Authors:

  • M. U. Ahmed;P. Funk

  • Affiliations:

  • School of innovation, design and engineering, Mälardalen University, SE-72123 Västerås, Sweden;School of innovation, design and engineering, Mälardalen University, SE-72123 Västerås, Sweden

  • Venue:
  • ISSPIT '11 Proceedings of the 2011 IEEE International Symposium on Signal Processing and Information Technology
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Rare cases are often interesting for health professionals, physicians, researchers and clinicians in order to reuse and disseminate experiences in healthcare. However, mining, i.e. identification of rare cases in electronic patient records, is non-trivial for information technology. This paper investigates a number of well-known clustering algorithms and finally applies a 2^nd order clustering approach by combining the Fuzzy C-means algorithm with the Hierarchical one. The approach was used to identify rare cases from 1572 patient cases in the domain of post-operative pain treatment. The results show that the approach enables the identification of rare cases in the domain of post-operative pain treatment and 18% of cases were identified as rare.