Paper Contents
Abstract
The enhancements made in medical care services have been inclining exponentially to improve their Quality of Service (QoS). The technological advancements have also increased the generation of data and the demand for extended services without deteriorating their quality. The medical care insurance sector is one of the widest applications that administers and reimburses medical amounts. One of the main scams that hinders the advancement of medical care services is scams detection in the insurance industry. Medical care scams mostly include entities like suppliers, beneficiaries, and subscribers. Existing procedures for supplier scams detection in health insurance have limitations such as false positives, a lack of real-time capabilities, and inadequate data integration. Static rule-based systems, high maintenance costs, and insufficient use of advanced technologies pose challenges. Privacy concerns, complexity in handling big data, and the need for better collaboration among stakeholders are additional issues. Overcoming these challenges requires ongoing innovation and the adoption of advanced technologies to enhance detection efficiency and efficiency. Accessing claim data gathered from suppliers sensibly is crucial for decision support. Hence, this investigation study aims to intelligently classify and detect scams behavior. The investigation study is conducted in three phases. The first phase presents a review of conventional ML-oriented techniques to identify the scope and problems in scams detection systems. Along with that, the validation of the defined problems and the significance of the accumulated dataset are also investigated. The second phase aims to detect the scams data using the designed MR-ISVM classifier. To begin the study, the capabilities of the data processing units are refined using a novel MapReduce system. Then, the ISVM is modulated using the Tanimoto Index (TI). The formulated MR_ISVM classifier has significantly reduced the effects of computational efforts and time with improved categorization efficiency. In an effort to understand the patterns of scams and build detection modules, a novel Supplier PFADS is finally provided. The Decision Score-based Bayesian Optimization (DS-BO) hyper parameter model is used to pick the best features for learning the scams conduct once the supplier's profile is constructed using the Relative Risk-based Map Reduce system (RR_MR). The categorization and detection of scams labels is done using Recurrent Neural Networks (RNNs). The recalling ability of the learning process in Recurrent Neural Network (RNN) is enhanced in combination with the DS-BO technique. At last, the fittest attributes under the chosen hyperparameters are taken into the input layer of RNNs, which exposes the suppliers scams. The experimental study is carried out on the medical care Insurance Scams Supplier dataset accumulated from Kaggle, a public repository. Each studys phase is explored and examined under the performance metrics of efficiency, precision, and recall through the confusion matrix strategy.
Copyright
Copyright © 2024 PARUL SAINI. This is an open access article distributed under the Creative Commons Attribution License.
