Feature-Based Transfer Learning Model for the Diagnosis of Breast Cancer

The chance of dying from breast cancer is disproportionately high, and the disease’s prevalence also makes it a substantial health problem. When breast cancer is correctly identified, the patient’s chances of obtaining therapy and surviving the disease improve dramatically. Trained medical personnel can use both designations interchangeably. The model was generated by utilising a deep recurrent neural network (RNN), which was then optimised using the Keras-Tuner approach. A method with the goal of enhancing the accuracy with which breast cancer is diagnosed. The following are some of the inputs that go into an optimised deep RNN, an output layer. There are five layers that are buried, five layers that are discarded, and an output layer. We invested a lot of time and effort into perfecting the discrete hidden layers. Typically, one of three major approaches to feature selection might be used, being used to achieve the task at hand, which requires selecting those qualities to identify the most important components of the database, which entails putting the qualities to use to complete the activity. Each of the five models covered here is an example of a well-known kind used in machine learning. The offered features have been added to and run by a range of models, including one with a modified deep RNN. Findings from this investigation suggested that the deep RNN with the given characteristics has reached the best feasible accuracy through the deployment of the univariate tuning technique. The outcomes were superior to those of rival models in their respective categories.