Millions of women all over the globe are afflicted with a life-threatening illness known as breast cancer, which is quite common. It is very necessary to make a prompt and correct diagnosis to provide appropriate therapy and achieve better results for the patient. In recent years, approaches based on Deep Learning (DL) that are used for data mining have shown promising results in automating the process of diagnosing breast cancer. This study reviews the use of DL methods, such as Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), in conjunction with data mining techniques, to improve breast cancer detection and classification. In this particular investigation, we read the earlier research work to conduct a review of it, and subsequently, based on that reading, we carry out a comparative analysis of several characteristics including accuracy, precision, recall, and F1-score. When using the Breast Cancer Wisconsin Dataset, the comparison study reveals that Umer et al., (2022) have the maximum accuracy, precision, recall and F1-score value, which is 100%. Conversely, Olatunji et al., (2023) have the lowest accuracy, precision, recall, and F1-score value while utilizing the published Kaggle datasets. Their respective values are 92.20%, 90%, 91.6% and 92.4%.