Application of Fuzzy Matter Element Model Based on Entropy Weight and GBDT Algorithm for Financial Fraud Behavior Identification

This study mainly focuses on the application of integrated entropy weight with GBDT algorithm on financial fraud identification. Based on fuzzy matter element analysis, this study used integrated entropy weight with GBDT algorithm to identify financial fraud behavior. Taking listed company as the research object, companies were classified as the fraud group and the non-fraud group according to the information disclosure evaluation results by Shenzhen Stock Exchange. The fuzzy matter element based on entropy weight was then used for further factor screening, and the fraud identification model was constructed based on GBDT algorithm. The overall recognition rate was 84.4
^%
; the identification rate of fraud group was 89.7%; and that of the non-fraud group was 83%. The inter-predictability and the accuracy of detection rate were significantly improved, and the results provided an efficient model to comprehend fraud behavior in financial markets.