Artificial Neural Network Based Ventilator Support System for Automatic Respiratory Assistance to Alzheimer's Disease Patients

After so many trial and error we got the mean square error of the trained model network as 0.019907 and R value as 0.852. To reduce over fitting hidden layer value must be taken less than twice the input layer size. New data set different from that used for modelling ANN was used for testing. Compared with the predictions of the physicians the artificial neural network model shows 86% accuracy CONCLUSION: Continuous oxygen saturation level monitoring and control is essential for Alzheimer's disease patient, since persistent respiratory diseases in these people can become a common reason for sudden death. For this a ANN model using MATLAB programming was developed to predict inspired oxygen amount which was given to the patient as life support. The suggested model output was found to be utilized as a suggestive method to assist clinicians because it displays an accuracy of over 80% when compared to the decisions made by physicians.