A robust and interpretable ensemble machine learning model for predicting healthcare insurance fraud

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Jan. 2, 2025

Healthcare insurance fraud imposes a significant financial burden on healthcare systems worldwide, with annual losses reaching billions of dollars. This study aims to improve detection accuracy using machine learning techniques. Our approach consists three key stages: data preprocessing, model training and integration, result analysis feature interpretation. Initially, we examined the dataset's characteristics employed embedded permutation methods test performance runtime single models under different sets, selecting minimal number features that could still achieve high performance. We then applied ensemble techniques, including Voting, Weighted, Stacking methods, combine compare their performances. Feature interpretation was achieved through partial dependence plots (PDP), SHAP, LIME, allowing us understand each feature's impact predictions. Finally, benchmarked our against existing studies evaluate its advantages limitations. The findings demonstrate improved offer insights into interpretability in this context.

Language: Английский

---

An advanced blockchain-based hyperledger fabric solution for tracing fraudulent claims in the healthcare industry

Decision Analytics Journal, Journal Year: 2024, Volume and Issue: 10, P. 100411 - 100411

Published: Feb. 2, 2024

Blockchain and Machine Learning (ML) are state-of-the-art technologies in the digital era. Developing countries have witnessed great transformation, one of key challenges during this development phase has been to create a platform that integrates health vitals citizens confidentially. The healthcare insurance industry is sector experienced significant number instances fraudulent claims mismanagement patient data. These issues often arise due inadequate technological integration an over-reliance on manual processes human intervention. relies multiple between end users initiate, maintain, close diverse policies. proposed model initially recommends suitable policy for newly admitted patients, but case existing objectives speed up transaction processing payment settlement securely using private blockchain. Collectively, these two possess potential revolutionize future. This study aims incorporate blockchain ML techniques like Support Vector (SVM) Random Forest Regression, which can differentiate legal medical records recommend personalized policies, streamline claim processing, ensure security sensitive information vital records. aim more patient-centric environment with data transparency. integrated framework results secure, adaptable, efficient ecosystem outperforms traditional methods, paving way future services.

Language: Английский

---

Graph data science: Applications and future

Elsevier eBooks, Journal Year: 2025, Volume and Issue: unknown, P. 227 - 243

Published: Jan. 1, 2025

---

Leveraging evolutionary algorithms with a dynamic weighted search space approach for fraud detection in healthcare insurance claims

Knowledge-Based Systems, Journal Year: 2025, Volume and Issue: unknown, P. 113436 - 113436

Published: April 1, 2025

Language: Английский

---

Machine Learning Models for Fraud Detection: A Comprehensive Review and Empirical Analysis

Deleted Journal, Journal Year: 2024, Volume and Issue: 20(3s), P. 1138 - 1149

Published: April 4, 2024

An in-depth familiarity with ML and DL models for fraud detection is essential due to the growing frequency complexity of fraudulent activity across many domains. Despite abundance research on subject, empirical analyses these models, especially in their real-time implementations, are typically lacking. This study fills that need by meticulously reviewing analysing developed detection. We draw attention shortcomings existing approaches, which crucial ever-changing field include problems recall, scalability, complexity, precision, accuracy. By evaluating various using measures, our evaluation method based a rigorous approach. Provided insights into practical consequences as well flexibility each model circumstances, they thoroughly assess performance. There ways this work will be useful; example, it help professionals choose best needs, improve academic knowledge practice, open door more studies concentrate developing better tools. comprehensive gives academics companies foundation better, effective scalable systems period digital security.

Language: Английский

---

Unraveling Patterns in Healthcare Fraud through Comprehensive Analysis

2022 9th International Conference on Computing for Sustainable Global Development (INDIACom), Journal Year: 2024, Volume and Issue: unknown, P. 585 - 591

Published: Feb. 28, 2024

This study addresses healthcare provider fraud in Medicare, employing advanced machine learning models on a diverse dataset to predict potential fraud. The goal is contribute insights for effective detection and mitigate its impact overall costs. Utilizing Logistic Regression, Random Forest, addressing class imbalance through SMOTE, we discern patterns behavior. Evaluation metrics, including confusion matrices, accuracy, sensitivity, specificity, Kappa values, AUC, F1-scores, comprehensively assess each model's performance. Findings highlight distinct behavior underscore SMOTE's effectiveness mitigating challenges. Comparative analysis discusses algorithm strengths weaknesses, offering real-world implementation, impacting prevention strategies insurance companies, providers, beneficiaries. In summary, this research makes noteworthy contribution the of fraud, strategies. incorporation SMOTE enhances model robustness, leading improved capabilities and, consequently, reducing

Language: Английский

---

Fraud Detection in Financial Transactions Using Deep Learning Approach: A Comparative Study

Published: May 24, 2024

Language: Английский

---

Real-time Insurance Fraud Detection using Reinforcement Learning

Published: Feb. 23, 2024

Reinforcement gaining knowledge of (RL) can revolutionize the field coverage fraud detection by supplying an extra dynamic and timely method than traditional strategies. RL works in a loop consecutive processing steps, so it adapt to changing environments wherein fraudulent sports arise dangers related them real-time. fashions analyze from beyond studies make choices on how excellent reply tries adjust variables used for detection. Such models examine diverse outside records sources, which include information reviews, public records, social media, detect any shifts inside detected activities. Furthermore, gives insurers additional layer safety that they will only have had entry after while relying entirely hit upon activities as quickly appear, simultaneously conventional strategies cannot. As such, take immediate effective movement lessen losses responding every case most appropriately. All all, high-quality opportunities businesses policyholders Overall, real-time insurance using reinforcement learning is game-changer industry. It not improves capabilities but also reduces costs caused This technology continuously evolving, becomes more advanced, become even combating fraud. Its potential use past sources approach its ability interest makes ideal businesses.

Language: Английский

---

Network analytics for insurance fraud detection: a critical case study

European Actuarial Journal, Journal Year: 2024, Volume and Issue: 14(3), P. 965 - 990

Published: May 14, 2024

Language: Английский

---

Enhancing healthcare data integrity: fraud detection using unsupervised learning techniques

International Journal of Computers and Applications, Journal Year: 2024, Volume and Issue: 46(11), P. 1006 - 1019

Published: Oct. 9, 2024

Data in healthcare forms the backbone of any treatment and decision-making for patients. However, data from a institution can sometimes be prone to abnormalities, thereby putting patient safety jeopardy. This paper points dire need reliable anomaly detection systems industry. It employs various unsupervised learning methods, including Isolation Forest, Local Outlier Factor (LOF), K-Nearest Neighbours (KNN), autoencoder models detecting abnormalities with better accuracy. Anomaly capabilities also allow health providers reduce risks provide some assurance integrity data, as these are more likely indicate unusual profiles or incorrect test results. LOF, KNN preliminary methods performing this work, Forest yielding best Then, that learn subtle variations complex patterns employed. seeks enhance terms accuracy reliability ensure quality safety.

Language: Английский

---