Theses and Dissertations Collection

Our research describes effective data mining based on the Named Entity Recognition (NER) technique for medical claims fraud / abuse detection system. Fraud and abuse in medical claims have become a major concern among health insurance companies in Saudi Arabia in recent years, not only by faking prices and numbers, but by assigning inaccurate ICD-10 codes to diseases that do not match the diagnosis of the claim. Handling medical claims is backbreaking manual work performed by a few medical experts who are responsible for approving, modifying, or denying applications for grants within a limited period of time from receipt. The proposed screening system uses a NER detection tool for each of the claims involved and classifies whether it is a fraud or not. Our research study has carried out an analysis for medical claim data from a private insurance company collected from different databases. Results from our fraud detection system show clearly that the number of suspicions of abuse is excessively high compared to the number of no suspicion of abuse.