A bayesian recurrent neural network for unsupervised pattern recognition in large incomplete data sets

Int J Neural Syst. 2005 Jun;15(3):207-22. doi: 10.1142/S0129065705000219.

Abstract

A recurrent neural network, modified to handle highly incomplete training data is described. Unsupervised pattern recognition is demonstrated in the WHO database of adverse drug reactions. Comparison is made to a well established method, AutoClass, and the performances of both methods is investigated on simulated data. The neural network method performs comparably to AutoClass in simulated data, and better than AutoClass in real world data. With its better scaling properties, the neural network is a promising tool for unsupervised pattern recognition in huge databases of incomplete observations.

MeSH terms

  • Algorithms
  • Antipsychotic Agents / adverse effects
  • Artificial Intelligence
  • Bayes Theorem*
  • Cluster Analysis
  • Creatine Kinase / blood
  • Data Interpretation, Statistical
  • Databases, Factual*
  • Humans
  • Mental Recall
  • Neural Networks, Computer*
  • Neuroleptic Malignant Syndrome / epidemiology
  • Pattern Recognition, Automated*
  • World Health Organization

Substances

  • Antipsychotic Agents
  • Creatine Kinase