Neural Network-Based Breast Cancer Histology Prediction in Moroccan Patients

Abstract

This paper studies the application of artificial intelligence, particularly artificial neural networks, in the classification of breast cancer within a Moroccan population, using data from the Hassan II Hospital in Agadir. The dataset includes comprehensive patient information such as age, environment, and various clinical analyses, to classify patients into two categories: IDC and Non-Idc breast cancer. Our developed neural network, a fully connected network with dropout regularization and using the Adam optimizer, was trained over 500 epochs and demonstrated promising performance, achieving an accuracy of 0.75 after 473 epochs. Finally, the model’s stability post-epoch 473 highlights its robustness, effectively mitigating overfitting and underfitting and the achieved accuracy of 0.75 indicates the model’s potential to provide valuable insights for medical professionals, aiding in breast cancer classification, treatment planning, and resource allocation in hospitals.