Outlier Detection

Outlier Detection on Credit Card Fraud and Boston datasets using statistical, supervised, and unsupervised methods. Statistical methods include non-parametric visualization techniques like boxplots and scatter plots, and multivariate parametric methods such as X2-test (Z-Score or F1-Score), and IQR score. From supervised methods, the isolation forest algorithm is performed. From unsupervised methods, the Local Outlier Factor (LOF) and the one Class SVM algorithms are tested.