Combinational neural logic system and its industrial application on hypoglycemia monitoring system

In this paper, a combinational neural logic network (NLN) with the neural-Logic-AND, -OR and -NOT gates is applied on the development of non-invasive hypoglycemia monitoring system. It is an alarm system which measured physiological parameters of electrocardiogram (ECG) signal and determine the onset of hypoglycemia by use of proposed NLN. Due to different nature of application, conventional neural networks (NNs) with common structure may not always guarantee the optimal solution. Based on knowledge of application, the proposed NLN is designed systematically in order to incorporate the characteristics of application into the structure of proposed network. The parameter of the proposed NLN will be trained by hybrid particle swarm optimization with wavelet mutation (HPSOWM). The proposed NLN will be practically analyzed using real data sets collected from 15 children (569 data sets) with Type 1 diabetes at the Department of Health, Government of Western Australia. By using the proposed method, the detection performance is enhanced. Compared with other conventional NNs, the proposed NLN gives better performance in terms of sensitivity and specificity.