Découverte et représentation des trajectoires de soins par analyse formelle de concepts

Nicolas Jay 1
1 ORPAILLEUR - Knowledge representation, reasonning
INRIA Lorraine, LORIA - Laboratoire Lorrain de Recherche en Informatique et ses Applications
Abstract : Chronic diseases are a major concern in public health. They require a multidisciplinary approach and chronic patients have multiple contacts with the healthcare system. In order to improve quality of life, quality and efficiency of healthcare, the care process of chronic patients must be assessed. This implies that the whole succession of health events of such patients should be monitored. From now on, in France, no information system was design to describe cross-institutional and long-term patient care trajectories. Moreover, there have been few attempts to classify and build a typology of patient trajectories. However, enormous amount of data are collected each year by several healthcare information systems. In this work, we propose a new approach to analyze and represent patient care trajectories from data that are originally collected for a different purpose. Our system is based on Formal Concept Analysis (FCA). FCA is a method of conceptual classification aiming at discovering natural relationships in a binary table and representing them in a concept lattice. We show the advantages of FCA for describing and understanding processes in the healthcare system and we give some perspectives for the exploration of social networks in general. We also study and compare two measures of interestingness of concepts : stability and support. The measures are computed in order to reduce the complexity of big concept lattices and filter the most relevant knowledge. Drawing a parallel with itemset mining and sequential itemset mining, we propose an unsupervised classification method with interesting properties for visualization and interpretation.
Document type :
Theses
Complete list of metadatas

Cited literature [14 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00585411
Contributor : Nicolas Jay <>
Submitted on : Tuesday, April 12, 2011 - 5:47:07 PM
Last modification on : Friday, May 24, 2019 - 10:14:05 AM
Long-term archiving on : Thursday, November 8, 2012 - 4:15:33 PM

Identifiers

  • HAL Id : tel-00585411, version 1

Collections

Citation

Nicolas Jay. Découverte et représentation des trajectoires de soins par analyse formelle de concepts. Interface homme-machine [cs.HC]. Université Henri Poincaré - Nancy I, 2008. Français. ⟨tel-00585411⟩

Share

Metrics

Record views

999

Files downloads

867