This repository stores the code for ML project - Breast Cancer Detection using the Breast Cancer Wisconcin (Diagnostic) data. The model uses logistic regression to classify whether a tumor is maligant or benign.
This model achieves a 95% accuracy in its training set, yet only achieves 68% and 65% in its validation and test set respectively. Clearly, it suffers from overfitting. Regularization and feature selection might be needed to improve the results.
After feature selection, this model achieves 94%, 87% and 95% in its training set, validation set and test set respectively.