<p>Unfortunately, there hasn’t been a lot of research on Breast Cancer (BC) detection and diagnosis that compares all of the well-known Machine Learning (ML)-based methods in one article. The paper studies 107 state-of-art research works within the timeline 2018 to 2023 and effectively pin points their flows to make further improvement. Review on BC detection and diagnosis using forty-eight well-known ML approaches and thirty-seven different feature selection methods have been observed in this study. The paper focused on different data pre-processing techniques including data balancing, outlier detection and removal, missing value handling, feature engineering and feature selection. The paper categorically describes the weakness of each research articles to make the scope of further improvement. It has been observed that in the vast majority of cases, ensemble algorithms outperform their conventional counterparts, and standalone classifiers. This paper discussed numerous metrics that were used to evaluate the performance of the BC detection and diagnosis systems in order to provide a clear understanding. The article also discussed some open issues, research gap, and future research direction of BC detection and diagnosis. Finally, using five benchmark datasets, the research examines the performance of models employing ten machine learning classifiers. The study revealed that the performance of ensemble classifiers was highly noteworthy. This work also serves as a useful reference for researching BC detection and diagnosis in both academic and applied contexts because it provides pertinent data on the most important machine learning based techniques used till date.</p>