Đang chuẩn bị liên kết để tải về tài liệu:
Management of fraud: Case of an Indian insurance company

Frauds in insurance are typically where a fraudster tries to gain undue benefit from the insurance contract by ignorance or wilful manipulation. Using the claims data in motor insurance obtained from a Mumbai based insurance company for the time period of 2010-2016, this study focuses on studying the pattern exhibited by those claims which have been rejected and accepted as well. | http afr. sciedupress.com Accounting and Finance Research Vol. 7 No. 3 2018 Management of Fraud Case of an Indian Insurance Company Sunita Mall1 Prasun Ghosh2 Parita Shah3 1 Assistant professor Statistics MICA Ahmedabad India 2 Business Analyst Hansa Cequity India 3 Senior Manager Actuarial TATA AIG General Insurance Company Ltd. India Correspondence Sunita Mall Assistant professor Statistics MICA Ahmedabad India Received May 12 2017 Accepted January 7 2018 Online Published April 29 2018 doi 10.5430 afr.v7n3p18 URL https doi.org 10.5430 afr.v7n3p18 Abstract Frauds in insurance are typically where a fraudster tries to gain undue benefit from the insurance contract by ignorance or wilful manipulation. Using the claims data in motor insurance obtained from a Mumbai based insurance company for the time period of 2010-2016 this study focuses on studying the pattern exhibited by those claims which have been rejected and accepted as well. The prime objective of the study is to identify the important or the significant triggers of fraud and predicting the fraudulent behaviour of the customers using the identified triggers in an existing algorithm. This study makes use of statistical techniques like logistic regression CHAID Chi Square Automatic Interaction Detection technique to identify the significant fraud triggers and to determine the probability of rejection acceptance of each claim coming in future respectively. Data mining techniques like decision tree and confusion matrix are used on the important parameters to find all possible combinations of these significant variables and the bucket for each combination. This study finds that variables like Seats Tonnage No Claim Bonus Type of Vehicle Gross Written Premium Sum Insured Discounts State Similarity and Previous Insurance details are found to be significant at 1 level of significance. The variables like Branch Code and Risk Types are found to be significant at 5 level of signify cance. The Gain chart depicts that .