Fused Hierarchical Neural Networks for Cardiovascular Disease Diagnosis

A fused hierarchical neural networks (FHNNs) is proposed for applications mainly related to diagnosis and fault detection. The benefit of such a model is well demonstrated by applying FHNNs for cardiovascular disease (CVD) diagnosis hierarchically using hemodynamic parameters (HDPs) derived from non-invasive sphygmogram (SPG). Patients' medical records with diagnostic results confirmed by doctors obtained from two hospitals in China are used to test and verify FHNNs. Variance analysis is used to categorize HDPs according to the importance/relevance based on their influence on discriminating diseases. Different neural networks structures are tested in diagnosing CVD so as to choose the optimal sub neural networks (sub-NNs) for the proposed FHNNs. Finally FHNNs with fused sub-NNs for CVD diagnosis is presented. The validity and effectiveness in the improvement of accuracy is witnessed clearly from the testing results.