چكيده به لاتين
Financial contracts, including electronic transactions, tenders, auctions, etc., are tools for the formation of value exchanges, and this definition includes a major part of economic exchanges. In the meantime, profiteers and offenders seek to obtain benefits in excess of their rights with various tricks, seek fraud and theft, and in all these processes, by using weak points and hidden angles, they do this more easily and with a lower probability of discovery. . Today, with the electronicization of many of these exchanges, we have good data on the details of the actions of each person in the process, which can be used to discover suspected violations and from this lead to identify violations and weak points of the process for more detailed audit and inspection. Due to the large volume of these data, artificial intelligence and machine learning methods are used, but one of the challenges for this work is the lack of recognition of violations, or in other words, the existence of a very small number of discovered violations compared to the total number of cases. The purpose of this applied research is to provide a A new and integrated approach to eliminate data imbalance using artificial generation of minority data is to improve the efficiency of machine learning classification models considering the conditions and limitations of the applied platform. In order to evaluate the method, the publicized data that have been used in other researches in this field have been used, and finally, by implementing it on the real data of the electronic tendering system of the Iranian headquarters, the implementation method has been shown in a real application. The results of the evaluation of the method showed a reduction in errors compared to conventional methods, as well as a high speed of implementation, as a result of which it can be implemented for big data.
