Cardiac health diagnosis using data fusion of cardiovascular and haemodynamic signals
Computer Methods and Programs in Biomedicine
| Hi-index | 0.00 |
The terrain location identification problem represents a very complicated learning task. Beside of learning from noisy and nondeterministic training data, the training task must learn from a very large size of training data, which may lead to lots of learning problems. A phenomenon called the fake convergence is observed in our implementation. In that case, the training process seemed to converge to a fixed error level, but the actual error is much higher than the converged one. In our study, a fuzzy hierarchical network is proposed to cope with the problem of large training data sets. With this fuzzy hierarchical structure, the learning process can become fast and errors are significantly reduced. Another issue is regarding about embedding domain knowledge into the learning structure of neural fuzzy networks. The idea is simple but effective. The proposed structure is called the fuzzy hierarchical data fusion network and its learning performance is significantly better than that of original fuzzy hierarchical networks. With the use of fuzzy hierarchical data fusion networks, errors indeed can converge and the system becomes practically applicable.