• Credit Card Transactions Data Adversarial Augmentation in the Frequency Domain

    Prabhakara Uyyala (see profile)
    Item Type:
    Permanent URL:
    With the quick improvement of web based business, an enormous measure of exchange information has been created, which might be utilized by lawbreakers for fake exchanges. In genuine application, not the entirety of the assets take episodes can be found by the inconsistency identification framework in the bank in light of the fact that the framework isn't sufficiently touchy. A lot of capital was taken by guilty parties each year; this makes a terrible impact in general society. In this way, it gets critical to distinguish such deceitful exchanges in the huge trade information. The development of science and innovation achieves the expansion of information volume, just as uncover the new use of AI, which gives scientists an approach to distinguish weakness. AI strategies are broadly utilized in the field of charge card misrepresentation recognition; however the unevenness between various classifications represents an obstruction in the learning errands. To reduce this sort of issue, in this paper a model dependent on recurrence space attributes be proposed, which is joined with a generative antagonistic organization (GAN) to expand minority class. Rather than the conventional antagonistic organization, which just creates ill-disposed examples from the preparation information itself, this methodology utilizes the recurrence space abundance highlights of the information to produce different preparing information that coordinates the pattern of information changes. Test results exhibit order execution is extensively improved and that over-fitting is mitigated when applying to different unique datasets. Additionally, outflank other existing cutting edge draws near.
    Published as:
    Journal article    
    Last Updated:
    7 months ago
    Share this:


    Item Name: pdf 292-ijaema-may-2021.pdf
      Download View in browser
    Activity: Downloads: 38