The research of Bayesian method from small sample of high-dimensional dataset in poison identification

In order to reduce the hazards of biochemical terrorist attacks to the countries and regions, Bayesian network model was used to identify poison according to the observed preliminary symptoms of the poisoning people. As the collected dataset had the characteristics of high-dimensional and small sample, we proposed a Bayesian network structure learning algorithm based on dataset extension, correction and feature selection, which could learn an effective Bayesian network structure from small sample of high-dimensional datasets. And it was verified to be correct and valid on the UCI datasets and biochemical poison dataset.