Breast cancer is a common malignancy among women and ultrasonography shines, unlike X-rays and CT scans which we use radiation to create images. In the current work there are few limitations accuracy less than 90% limited classes used and traditional methods used for image processing. It is a crucial regular examination for the identification of breast cancer. Still, breast cancer diagnosis accuracy is not very high. Next, an accurate diagnosis using an image from a breast ultrasound (BUS) would be important for numerous computer-aided diagnostic learnings. There are techniques that can cause breast cancer lesion classification and diagnosis. This project is mainly focused on the classification of breast cancer by implementing the hybrid algorithm of XGBoost and linear regression model for the classification of classes. To achieve this loading and preprocessing, we must apply rules and methods of the deep learning concept and the dataset for the feature extraction of the classes, saving it as a Numpy file, then applying the hybrid algorithm of XGBoost, linear regression for the classification purpose, and at last finding the accuracy, precision, and recall of the model for the evaluation. This hybrid model and Convolutional Neural Network (CNN) get an accuracy of almost 97% for the classification.