Breast Cancer Prognosis Based on Machine Learning Model

Early detection is crucial for prompt and effective treatment of breast cancer, which is a serious health risk. Breast Cancer in particular, frequently held cancer among women fraternity. Breast cancer is a condition where the cells of mammary gland grow excessively and abnormally, forming a mass known as a tumor. Increasing numbers of cancer diagnosis and death due to cancer make it a most important medical problem. Though, in initial stage, the cancer detection process may avoid many deaths and critical cancer related issues. In order to create the prognostic classification techniques, this may be utilized for forecasting results for specific cancer patients. For this purpose, increasing number of Machine Learning techniques has played a vital role. In this work, we present the prognosis of breast cancer based upon machine learning and image processing techniques. The most common method for detecting breast cancer is mammography. Wisconsin Diagnostic Breast Cancer dataset (WDBC) is the widely used cancer dataset for early cancer predictions. Some frequently used techniques with cancer datasets are Naive Bayes, Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbour, Support Vector Machine along with Extreme Gradient Boosting has utilized for designing the model. Here we have focused machine learning technique, as SVM has secured the accuracy of 99.10%. In this article, we have also studied the outcome of the above techniques and compared their performance. Moreover, our experiment with WDBC shows that SVM outperforms other techniques.