RFSVM: A Novel Classification Technique for Breast Cancer Diagnosis

Abstract

Cancer is a disease, which develops, in human body due to gene mutation. Due to various factor cells turn into cancerous cell and grow rapidly while damaging normal cells. Many women get affected by breast cancer, which might even cause death if not treated at early stage. Early detection of breast cancer is highly important to increase the survival rate. Machine learning methods and technologies are making it possible to classify and detect the class in an accurate manner. Among other classifiers, random forest and support vector machine are two classifiers that have a good classification power. In this, research a combination of these two classifier i.e. Random Forest and Support Vector Machine (RFSVM) is proposed for early diagnosis of breast cancer cell using Wisconsin Breast Cancer Dataset (WBCD). Using different train-test data ratio experiments are performed and an average of more than 98percentage accuracy is achieved using this hybrid classifier. This paper overcomes the over-fitting problem of random forest and the need of tuning the parameters of Support Vector Machine. Even with limited data available, the classifier tunes its parameters so well to give a highly accurate result.

Breast Cancer Classification Approaches - A Comparative Analysis

“…The most accurate models are LR, SVC, and DT, according to the results.. Seventy percent of the data is used for training, while the other thirty percent is used for testing. Data is divided in a 7:3 (70:30) ratio into training and testing portions because cross-validation is not a possibility here [8][9], the algorithms for classifying data performed admirably on the classification task.…”