Advertisement

Hierarchical Dirichlét Learning – Filling in the Thin Spots in a Database

  • Steen Andreassen
  • Brian Kristensen
  • Alina Zalounina
  • Leonard Leibovici
  • Uwe Frank
  • Henrik C. Schønheyder
Conference paper
  • 442 Downloads
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2780)

Abstract

Estimation of probabilities by classical maximum likelihood estimators can give unreliable results when the number of cases is small. A Bayesian approach, where prior probabilities with Dirichlet distributions are used to temper the estimates, can reduce the variance enough to make the estimates useful. This is demonstrated by using this approach to estimate mortalities of severe infections from different sites, lungs, skin urinary tract, etc. The prior probabilities are provided in a hierarchical way, i.e. by deriving them from the same database, but without distinguishing between different sites of infection.

Keywords

Antibiotic Therapy Classical Estimator Dirichlet Distribution Empirical Antibiotic Therapy Average Mortality 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Andreassen, S., Leibovici, L., Schønheyder, H.C., Kristensen, B., Riekehr, C., Kjær, A.G., Olesen, K.G.: A decision theoretic approach to empirical treatment of bacteraemia originating from the urinary tract. In: Horn, W., Shahar, Y., Lindberg, G., Andreassen, S., Wyatt, J.C. (eds.) AIMDM 1999. LNCS (LNAI), vol. 1620, pp. 197–206. Springer, Heidelberg (1999)CrossRefGoogle Scholar
  2. 2.
    Andreassen, S., Riekehr, C., Kristensen, B., Schønheyder, H.C., Leibovici, L.: Using probabilistic and decision-theoretic methods in treatment and prognosis modeling. Artif. Intell. Med. 15, 121–134 (1999)CrossRefGoogle Scholar
  3. 3.
    Heckerman, D.: A tutorial learning with Bayesian networks. Technical Report MSR-TR-95- 06. Redmond, U.S.A (1995)Google Scholar
  4. 4.
    Kristensen, B., Andreassen, S., Leibovici, L., Riekehr, C., Kjær, A.G., Schønheyder, H.C.: Empirical treatment of bacteraemic urinary tract infection. Evaluation of a decision support system. Dan. Med. Bull. 46, 349–353 (1999)Google Scholar
  5. 5.
    Kristensen, B., Larsen, S., Schønheyder, H.C., Leibovici, L., Paul, M., Frank, U., Andreassen, S.: A decision support system (DSS) for antibiotic treatment improves empirical treatment and reduces costs. In: Proceedings of 41st Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), Chicago, Illinois, USA, December 16-19, vol. 476 (2001)Google Scholar
  6. 6.
    Leibovici, L., Shraga, I., Andreassen, S.: How do you choose antibiotic treatment? Br. Med. J. 318, 1614–1616 (1999)Google Scholar
  7. 7.
    Leibovici, L., Fishman, M., Schønheyder, H.C., Riekehr, C., Kristensen, B., Shraga, I., Andreassen, S.: A causal probabilistic network for optimal treatment of bacterial infections. IEEE Trans. Knowledge and Data Eng. 12, 517–528 (2000)CrossRefGoogle Scholar
  8. 8.
    Spiegelhalter, D.J., Myles, J.P., Jones, D.R., Abrams, K.R.: Bayesian methods in health technology assessment: a review. Health Tech. Assess. 4(38) (2000)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • Steen Andreassen
    • 1
  • Brian Kristensen
    • 2
  • Alina Zalounina
    • 1
  • Leonard Leibovici
    • 3
  • Uwe Frank
    • 4
  • Henrik C. Schønheyder
    • 5
  1. 1.Center for Model-based Medical Decision SupportAalborg UniversityAalborg EastDenmark
  2. 2.Department of Clinical MicrobiologyAarhus University HospitalÅrhus CDenmark
  3. 3.Department of MedicinePetah-TiqvaIsrael
  4. 4.Institut für Umweltmedizin und KrankenhaushygieneFreiburg Universiy HospitalFreiburgGermany
  5. 5.Department of Clinical MicrobiologyAalborg HospitalAalborgDenmark

Personalised recommendations