Breast cancer, which poses a danger to women's health and is a major cause of mortality, is a major focus of medical image analysis. Digital mammography can greatly increase the accuracy of disease detection by providing an early and accurate diagnosis of breast cancer. The approach uses transfer learning from the pre-trained ResNet-50 CNN on ImageNet to train and classify the breast dataset into benign or cancerous categories. It can be difficult to understand the intricate mammographic pictures, though. Convolutional neural networks (CNNs) have demonstrated promising results for image classification applications, such as breast cancer diagnosis, in recent years. In this regard, ResNet50 and AlexNet are two well-known CNN architectures that have been applied to the detection of breast cancer using mammographic images. In this study, performance evaluations of the machine learning algorithm such as Alexnet, and Restnet 50 using a dataset of breast cancer diagnoses were done. The effectiveness of the strategy was also tested to see how much AdaBoost's usage of weak learners affect. The primary goal is to assess the algorithm’s accuracy, precision, sensitivity, and specificity in terms of its effectiveness and efficiency in detecting data.