Answers are grounded in this paper and cite their sources.
The design and operation of blood pumps are very complex and expensive to save lives of patients suffering from heart ailments. Machine learning approaches may be those better affordable ones which can be used to enhance understanding of correlation between operating parameters and performance of the pumps that is between pressure rise in blood pumps and clearance, blade height and rotational speed of the pump. Four types of machine learning algorithms, that is multiple linear regression, polynomial regression, artificial neural network and random forest regression have been employed in the current study to predict pressure rise. Performance evaluation metrics have shown that machine learning approaches are efficiently able to predict the pressure rise in blood pumps using its operational parameters. Among all of the models used, the ANN (3‐12‐12‐1) model has shown the highest accuracy of 99.9% with RELU activation function for input and hidden layers, linear activation for the output layer using ADAMAX as an optimiser. The ANN model is found to be accurate with 0.0012 MSE, 0.023 MAE and 0.999 R 2 , which are significantly lower while predicting the pressure rise in the blood pump. The ANN model is also found to predict the physical behaviour of the blood pump efficiently.
No takes yet. Share an insight, caveat, or question.
Kumar et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: