Committee Machines with Ensembles of Multilayer Perceptron for the Support of Diagnosis of Heart Diseases

In this paper, committee machines (CM) based on ensembles of multilayer perceptron (MLP) are proposed to differentiate the diagnosis of five major heart diseases. In particular, two ensemble methods - random subspace and bagging, are considered. A heart disease database with 352 cases is used to evaluate the performance of the proposed schemes, and the experimental results show that both of the committee machines with ensembles of MLP outperforms the single MLP. Moreover, the committee machine based on the random subspace ensembles of MLP outperforms the one based on bagging ensembles. The obtained high classification accuracy proves the usefulness of the proposed committee machines in support of clinical diagnosis decision process