Credit card fake detection has raised unique challenges due to the streaming, imbalanced, and non-stationary nature of the data that has been transacted. It had additionally included an active learning step, since the labeling (fake or genuine) use of a subset on transactions is obtained in near-real time through human investigators contacted the cardholders. In this paper, the Hidden Markov Model (HMM) algorithm has been used for sequence of Credit card operations for transaction processing and the fake can be detected by using the fake detection model during transaction processing. HMM, Fake detection model and image process had played an imperative role in the detection of credit card fake in online transactions. In fake detection, most challenging is a data problem, due to two major reasons – first, the profiles of cardholders are normal and fake lent behaviors changed constantly and secondly, credit card fake data sets are highly changed its position. Using fake detection (FD) algorithm the performance of detection in credit card transactions had highly affected by the sampling approach on dataset, selection of HMM, Fake detection model. Using fake detection (FD) algorithm an image technique had been used. A reliable augmentation of the target scarce population of fakes are important considering issues such as labeling cost; algorithm HMM, fake detection and outlines in the data streamed source. We have approached several scenarios which showed the feasibility of improving detection capabilities evaluated by means of receiver operating characteristic (ROC) curves and several key performance indicators (KPI) commonly used in financial business.
Real-Time Optical and Electronic Sensing with a β-Amino Enone Linked, Triazine-Containing 2D Covalent Organic Framework
