Enhancing Breast Cancer Detection through Machine Learning: A Comparative Analysis of Logistic Regression and Linear Discriminant Analysis
Carlo V. Belvis, Claire Ann R. Javierto, Matilde Santos, Crizelle R. Datinggaling, Christian Van R. Rivera · 2023
This paper intends to analyze classification models that can predict whether breast cancer is cancerous or noncancerous through the attributes of the selected dataset with the application of machine learning algorithms. The high dimensional Breast Cancer Wisconsin (Diagnostic) Dataset is reduced through principal component analysis (PCA). Six major components were extracted with a cumulative variance of 91.063%. The selected classifier models that were evaluated were logistic regression and linear discriminant analysis (LDA). After training and testing, the model’s performance was evaluated using a confusion matrix where metrics such as accuracy, precision, recall, and F1-score were calculated. Experimental results obtained a commendable prediction with 99.7% accuracy for logistic regression and 95.4% accuracy for LDA. The LDA model was observed to have greater misclassified diagnoses than logistic regression. Thus, it is inferred that in binary classification problems like breast cancer detection, logistic regression produces more promising predictions.