An estimated 30 million new cases of breast cancer are expected to occur by 2030, making it a serious global health concern. Although curative actions are essential, improving both diagnostic and preventative methods are equally crucial in throughout reducing the severity of this crisis. This study introduces a machine learning-based method for classifying breast cancer using the Invasive Ductal Carcinoma (IDC) dataset, which is used as a standard to assess how well different ML algorithms differentiate between the Non-IDC and IDC class labels. We evaluate how well a number of machine learning algorithms such as non-liner Support Vector Machines (SVMs), Random Forests (RFs), and Back Propagation Neural Networks (BPNNs) perform in correctly classifying images of breast cancer. Conventional techniques usually use image patch extraction, which can be computationally demanding, for model training and testing. Despite these difficulties, our method produced results that were significantly better than those previously published. The ANN classifier specifically achieved sensitivity, specificity, and AUC-ROC of 92%, 91%, and 93%, respectively, and an average image-level accuracy of roughly 91%. The experimentation results prove that the ML-based approaches have the ability to extract some relevant information from the microscopic images for diagnosing breast cancer. The results also proves that the proposed methodology outperforms traditional diagnostic methods using accuracy and other measures.