### Roland Orre, Andrew Bate and Marie Lindquist.

### Abstract

The data mining task we are interrested in is to find
associations between variables in a large database.
The method we have earlier proposed to find outstanding associations
is to compare
estimated frequencies of combinations of variables with the
frequencies that would be predicted assuming there were no
dependencies.
The method we now propose use the same strategy as
an efficient way of finding complex
dependencies, i.e. certain combinations of explanatory
variables, mainly medical drugs, which may be highly associated
with certain outcome events or combinations of
adverse drug reactions (ADRs). Such combinations of ADRs may
also be recognized as syndromes.

The method we use for data mining is an artificial neural network
architecture denoted Bayesian Confidence Propagation Neural Network
(BCPNN). To decide whether the joint probabilities of events are
different from what would follow from the independence assumption, the
*"information component"* log(*P*_{ij} /(P_{i}P_{j})), which is
a weight in the BCPNN, and its variance plays a crucial role. We also
suggest how this method might be used in combination with
stochastic~EM to analyse conditioned dependencies also between real
valued variables, e.g. to consider the amount of each drug taken.

**Bayesian Neural Networks used to Find Adverse Drug Combinations and Drug Related Syndromes**

Roland Orre, Andrew Bate and Marie Lindquist.

In *Proc. of the ANNIMAB-1 Conference*, pages 215-220, Gothenburg, Sweden, April 2000. Springer.

(PDF)
Bayesian Neural Networks used to Find Adverse Drug Combinations and Drug Related Syndromes

Roland Orre
Last modified: Mon May 31 13:27:53 CEST 2010