A Comparative Analysis of Selected Machine Learning Classifiers for Early Detection of Breast Cancer

Abstract

Breast cancer is the principal causes of death among women in Nigeria and the world at large. The high prevalence of breast cancer in women has rapidly increased in recent decades. X-ray mammography and magnetic resonance imaging (MRI) are currently the major methods used to diagnose cancer in Nigeria. Machine learning techniques are the trendiest tool of the 21st century for discovery where intelligent approaches are utilized to detect breast cancer classification patterns. A technique for excellently recognizing and forecasting early cancer of the breast is needed by specialists for an unequivocal detection. This work provides a comparative analysis of selected Machine Learning classifiers for early detection of breast cancer. Locally dataset was obtained from Ekiti State University Teaching Hospital. Statistical approaches were employed to clean the data by assigning the missing value based on the neighboring feature values, in contrast to other methods that exclude all cases. This study examined three of Machine Learning techniques - Support Vector Machine (SVM), Logistic Regression (LR), and Naive Bayes (NB) models on the basis of which we idetified the best predictive model. Accuracy, Precision, Recall, and F1 Score metrics were implemented for evaluating each model’s performance using python programing language. The findings of this study demonstrate that the suggested SVM-based model has greater promise for breast cancer classification than the LR and NB classifiers yielding a 96% accuracy rate and suggested that the mechanism is effective than traditional techniques when assessed using several indicators of effectiveness to detect breast cancer.