Dynamic modeling of bayesian classifier based on general information theory
Hailong Zhang · Journal of Jilin University · 2009
By reasoning the conditional independence and dependence between attribute values based on General Information theory,redundant attributes are removed naturally and a submodel is constructed for each test sample rather that the whole sample space.The joint probability density and conditional probability is estimated based on marginal computation.The experimental study on the UCI data set shows that,this algorithm can describe the marginal dependency of mixed-mode data more intuitively and accurately.