Performance Comparison of SVM and Logistic Regression in Breast Cancer Diagnosis Using the WDBC Dataset

Background Breast cancer remains a serious worldwide public health problem that requires precise diagnostic tools. We compared two machine learning algorithms, Support Vector Machine (SVM) and Logistic Regression( LR) for the diagnostic task of breast cancer in this work. This work uses the Wisconsin Diagnostic Breast Cancer (WDBC) dataset, containing 569 examples split into different groups and each one identified by 30 numeric features which have already been computed regarding digital images of FNA tests. The dataset has two class of diagnostic labels: benign and malignant. Here, after putting both algorithms to the test we saw that SVM and LR had very good recall, accuracy, precision f1 score and ROC AUC Score. Especially Logistic Regression slightly performed better in a lot of important criteria. A full comparison was made to make up for what these models lacked in other measures were used. When looking at the ratio of true predictions to all predictions generated, both model experts have highest underscore for LR. The precision, that is the percentage of true positives among predicted positives) was also higher in LR. The recall measure, which denotes how many true positives are successfully identified, also showed a similar performance for both models. F1 measure which is a harmonic mean of recall and precision a little increase for LR. Finally, ROC AUC score quantified the overall discriminative capacity of classes that was slightly higher for LR.