Chronicles mining in a database of drugs exposures

Abstract : Pharmaco-epidemiology is the study of uses and effects health products on population. A new approach consists in using available data\-bases of care pathways to make studies. In such context, knowledge discovery approaches becomes a mandatory to support clinicians to formulate new hypotheses. Care-pathways are based on timestamped events and can be complex. Chronicles is one of the temporal patterns that can capture such complex temporal behaviors. In this paper, we adapt the existing chronicles mining algorithms in order to mine care-pathway which contains drugs exposures and medical problems, like hospitalizations. We present our method to extract all the frequent chronicles and we show the difficulties we encountered. Finally, we present our first results and our perspectives.
Type de document :
Communication dans un congrès
ECML Doctoral consortium, Sep 2015, Porto, Portugal. Proceedings of the ECML Doctoral consortium. <http://www.ecmlpkdd2015.org/submission/doctoral-consortium>
Liste complète des métadonnées

https://hal.inria.fr/hal-01184100
Contributeur : Thomas Guyet <>
Soumis le : mercredi 12 août 2015 - 17:19:30
Dernière modification le : samedi 18 février 2017 - 01:19:38

Identifiants

  • HAL Id : hal-01184100, version 1

Citation

Yann Dauxais, David Gross-Amblard, Thomas Guyet, André Happe. Chronicles mining in a database of drugs exposures. ECML Doctoral consortium, Sep 2015, Porto, Portugal. Proceedings of the ECML Doctoral consortium. <http://www.ecmlpkdd2015.org/submission/doctoral-consortium>. <hal-01184100>

Partager

Métriques

Consultations de la notice

434