Post-Surgical Survival Forecasting of Breast Cancer Patient: A Novel Approach

Cancer is the most common death causing disease and breast cancer is the deadliest cancer affecting women universally. Survivability forecast of a patient after breast cancer surgery become most challenging and difficult task to reduce the death rate. This survival prediction is associated with the life of a woman hence efficient algorithms must be used for the prediction purpose. Many prediction algorithms have been published in the field of post-surgical survival (PSS) prediction during past three decades. These approaches involve statistical or machine learning methods to forecast the survival of breast cancer patients; advised for lumpectomy/mastectomy. In this research work, a novel prediction approach is proposed using support vector machine-communication efficient distributed dual coordinate ascent (SVM-CoCoA). The authenticity of the proposed model is done with an experiment performed on the simple build tool (SBT) and Apache Flink. For experiment the Haberman's Survival dataset is used whereas data imbalance associated with this dataset is eliminated using the Synthetic Minority Oversampling Technique (SMOTE) to overcome the biased outcome.