ADS Capstone Chronicles Revised

14

Figure 7 Heatmap of Top Feature Contribution to Each Principal Component

Bagging Classifier, and Quadratic Discriminant Analysis were compared. 4.4.1 Baseline Model Selection - Logistic Regression Logistic regression was chosen as the baseline model due to its simplicity and effectiveness in binary classification tasks ( LogisticRegression , n.d.). Multiple data preprocessing approaches were evaluated to determine the optimal data representation for the Logistic regression model. The dataframes that were considered included

4.4 Modeling Various machine learning models were employed to detect potential fraud, waste, and abuse within the healthcare system using the Market Saturation and Utilization State-County dataset. Logistic Regression serves as the baseline model, against which the performance of more complex models like Ridge Logistic Regression, Lasso Logistic Regression, XGBoost, Neural Network Adaboost, SVM with Kernel Trick,

138