This paper provides an overview of the advancements and applications of machine learning (ML) techniques in breast cancer research. It discusses the use of ML algorithms for classifying cancers, predicting survival status, and forecasting recurrence. The survey reveals that machine learning approaches were employed for cancer classification, prognosis, and the identification of prognostic factors, while deep learning approaches were used for predicting patient survival subtypes, identifying concealed features, and predicting metastasis. Artificial neural network (ANN) approaches are effective in classifying cancer types, while logistic regression and K-nearest neighbours (KNN) perform poorly on certain datasets. SVM shows moderate performance. Feature ranking and filtering improve results and aid in prognosis. Random forest adapts well to multiple datasets. ML techniques are also applied to identify prognostic factors for breast cancer survival. It is found that genomic factors combined with clinicopathological data aid in early detection and recurrence prediction. ML models help identify genes responsible for recurrence. Various techniques, including survival analysis and mechanistic models, are employed. CNN shows promise in predicting breast cancer from microscopic images. LDA and random forest perform well in predicting metastasis. The survey emphasizes the need for validation on external datasets and the inclusion of larger datasets and varied data sources for improved accuracy.