In recent times, breast cancer has become one of the major health concerns for women which can have a severe impact on the overall well-being of the affected and may even result in death. Women, being an essential part of today’s world, require a top-notch healthcare system that can ensure their overall well-being and address even sensitive issues like breast cancer. So, it is essential to develop a breast cancer detection system that can detect breast cancer effectively and discriminate between dangerous malignant and benign tumors incorporating advanced machine learning techniques. Notably, promising results have been shown by algorithms like K-Nearest Neighbors, Decision Trees, and Naïve Bayes. Nevertheless, previous studies have only examined the model’s overall performance. The models contain no information regarding the multiclass performances of the evaluation matrices. To address these gaps, this study aims to identify the most efficient algorithm for classifying and predicting breast cancer. The analysis comprises a dataset that encompasses two distinct kinds of breast tumors: Malignant and Benign. This study evaluates the performance of the three mentioned methods by conducting thorough experiments employing k-fold cross-validation. Furthermore, the multiclass classification of various tumors showed superior performance compared to the K-Nearest Neighbors technique and other algorithms. This approach for evaluating the performance of several categories would provide important findings and an in-depth understanding of the overall model's performance which can be useful for future research.