The role of using data mining methods in enhancing the effectiveness of management fraud detection "An applied study using the particle swarm algorithm"

Abstract

The research aimed to test the role of using data mining methods in enhancing the effectiveness of detecting fraud in financial statements in the Syrian business environment, by applying it to the industrial and agricultural companies listed on the Damascus Securities Exchange, which are (3) listed companies for nine financial periods; From (2012) to (2020) with a total number of (27) observations. To reach the aim of the research, the Beneish M Score model was applied as a traditional method for detecting fraud and classifying the companies under study into fraudulent companies (containing fraud) and non-fraudulent companies (containing fraud). Then, the results were compared with the result of applying data mining method using the Particle Swarm Optimization (PSO) Algorithm supported by the Support Vector machines (SVM) algorithm.The research found a set of results, The most important of them are: There is a difference in the results between the auditor's use of traditional auditing methods in detecting fraud and his use of data mining methods with a difference of (83.33%) in the light of the study sample. The research also showed the superiority of data mining methods

 (Particle Swarm Optimization (PSO) Algorithm with the Support Vector machines (SVM) algorithm) in predicting management fraud over traditional statistical models that are used to predict fraud cases.