Publication Date: 2023/08/26
Abstract: Breast cancer is considered one of the biggest killers in women globally. The major reason of mortality is the reason that cancer is diagnosed at later stages. Objectively, this study is conducted to compare evaluation metrics of 6 ML models such as Naïve Bayes, k-Nearest Neighborhood (K-NN’s), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM) and Logistic Regression (LR) on Wisconsin Breast Cancer (BC) Dataset. WEKA tool has been used to calculate the performance evaluation of these supervised ML algorithms. The literature shows that the Weka tool has been widely used in various data mining problems. The results clearly show that two models have achieved better accuracy, recall and other performance metrics in order to identify risk of breast cancer in women. These two models are K-NNs and Random Forest. In conclusion, these supervised classifiers have been trained to detect malignant and benign cells. In the future, this study may be extended for BC classification on medical images on larger dataset in order to diagnose cancer at early stages.
Keywords: Machine Learning Algorithms, Breast Cancer, WEKA, ML Classifiers.
DOI: https://doi.org/10.5281/zenodo.8285931
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT23AUG664.pdf
REFERENCES