To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
Fuzzy-NNARX based Tool for Monitoring and Predicting Patients Conditions using Selected Vital Signs |
|
Author |
Cecilia H. Vallejos de Schatz, Fabio K. Schneider, Paulo J. Abatti, Julio C. Nievola |
| Citation |
Vol. 15 No. 1 pp. 113-122 |
|
Abstract |
In this paper, an artificial intelligent tool is proposed using fuzzy logic (FL) and recurrent neural networks (RNN) for definition and forecast of patient’s clinical condition. The fuzzy logic-based proposed first phase of the tool permits the analysis of the current state of the patient, which allows the training of the artificial neural network. In the second phase, two Elman networks Multi Input Single Output (MISO), two Elman networks Multi Input Multi Output (MIMO), as well as two Auto-Regressive Neural Networks with eXogenous inputs (NNARX) are evaluated with and without pruning. The fuzzy model agrees 99.76% with the answers given by the experts. After analyzing the six proposed networks, it was verified that the pruned NNARX model can offer the highest overall accuracy (OA) of 99.82%, whereas the others show a decrease of up to 35%. Finally, to implement the smart software of this paper, the best scenario was found to be the Fuzzy-NNARX solution where an OA of 99.25%, a sensivity of 99.62%, and a specificity of 99.83% was obtained by utilizing unseen data from thirty patients. More tests made with higher prediction periods (10, 30 and 60 seconds) demonstrate a slight decrease in the OA reaching up 94.58%. Nevertheless, the OA still remained over 94%. For the data used in this work, NARX networks capture the dynamics of nonlinear dynamic systems much better than Elman networks. Results demonstrate that the Fuzzy-NNARX model proposed has a very good performance in predicting the patient conditions, and it is a useful tool for preventive medicine for chronic patients |
|
Keywords |
Artificial Neural Network, Fuzzy Logic, Recurrent Neural Network, Neural Network Auto-Regressive model with eXogenous inputs, NNARX, Elman |
|
URL |
