An Empirical Study of Medical Diagnosis Using Deep Learning
Chandra Prakash Lora, Asha Rajiv, Jyotirmaya Sahoo · 2024
this paper provides an empirical have a look at investigating the effectiveness of deep mastering (DL) for medical prognosis. The authors completed a scientific look at on two one-of-a-kind scientific issues: blood cellular and breast cancer analysis. For every trouble, the authors explored the performance of 3 distinctive DL models, particularly convolutional neural networks (CNNs), deep notion networks (DBNs), and recurrent neural networks (RNNs). Furthermore, the authors examine the accuracy of the DL fashions to a baseline non-DL technique. Experiments on a benchmark dataset show that DL models reap higher accuracy and quicker convergence than the non-DL method, accordingly supplying a sturdy indication that the use of DL is powerful for scientific diagnosis. The authors inspect additionally exclusive combinations of DL fashions with specific feature choice strategies, as well as specific hyperparameter tuning techniques to enhance accuracy and pace of diagnosis. Moreover, the authors examine the steadiness of DL fashions inside the presence of noisy information. The results of this study exhibit that DL is a powerful tool for clinical prognosis and ought to be similarly explored.