Claim fraud detection is a common problem in insurance industries. Machine learning algorithms are handy in detecting fraud claims. Sample dataset downloded from https://github.com/mwitiderrick/insurancedata/blob/master/insurance_claims.csv
The dataset is evaluated using cross-validated score, ROC Curve and AUC. The predictive power of each model expressed by ROC curves. For instance, Linear Discriminant Analysis and XGBOOST model has higher probability of accurate prediction of correct class members, and gaining high level of accuracy prediction probability as compared to Random Forest, KNN, Naive Bayes, Neural Network and SVM models.