<![CDATA[This research presents the development process of a machine learning neural network model for predicting hemodynamic changes in patients after premedication, aiming to enhance the performance of medical interventions. The model was constructed using 3055 patients’ data who underwent premedication processes. The developed neural network model has an architecture consisting of 10 nodes in the input layer, 10 nodes in the hidden layer, and 3 nodes in the output layer. The evaluation results of the model indicate an overall accuracy of 85%. The precision values are high for normal class predictions at 0.85 and for hypertension class predictions at 0.81 with corresponding recalls of 1 (high) and 0.6 (moderate), respectively. However, predictions for the hypotension class still have a low precision of 0.6 and a recall of 0.04 (very low) due to the significantly lower number of samples in the hypotension class compared to the normal and hypertension classes. While testing with new data, the model has successfully predicted whether patients will experience hemodynamic pressure changes. It is expected that this model can contribute to improving the performance of medical interventions, thereby minimizing undesirable hemodynamic pressure changes.]]>
Copyrights © 2024
