Cybersecurity enhancement to detect credit card frauds in health care using new machine learning strategies

Abstract

As the usage of credit cards has become more common in health care applications of everyday life, banks have found it very difficult to detect credit card fraud (CCF) systematically. The fraudulent activities should be identified and detected using new techniques. As a result, machine learning (ML) can help detect CCF in transactions while reducing the strain on financial institutions. This research aims to improve cybersecurity measures by detecting fraudulent transactions in datasets. The new classifier strategies cluster and classifier-based decision tree (CCDT), cluster and classifier-based logistic regression (CCLR), and cluster and classifier-based random forest (CCRF) are modeled in this research. The proposed strategies are applied to detect fraudulent health care activities. This research performed the preprocessing through the feature extraction, sampling, and transformation stages, and the proposed classifiers are simulated, and the results are analyzed. The significant results expected range of the proposed classifiers over the other methods are accuracy—(99.95%, 99.97%), precision—(99.96%, 99.98%), sensitivity—(99.9%, 100%), specificity—(99.8%, 100%). The parameters \(\mu \), location, the binary variable, cluster size, and decision tree sampling observations affect the classifiers' performance. CCRF and CCLR obtain the expected significant results than other existing methods.