An effective fraud detection using competitive swarm optimization based deep neural network
T. Karthikeyan, M. Govindarajan, V. Vijayakumar
Abstract
Frauds have no persistent patterns. They constantly alter their behaviour, necessitating the use of unsupervised learning. Fraudsters gain access to latest technology that enables them to commit scams via internet transactions. Consumers' habitual behaviour is assumed by fraudsters, and fraud trends very quickly. Fraud detection systems must utilize unsupervised learning to identify online payments since some fraudsters initially use online channels before switching towards other methods. The goals of this research are to construct a deep convolutional neural network model to detect anomalies from regular patterns produced by competitive swarm optimization, with a particular emphasis on fraud scenarios that cannot be detected using prior records or supervised learning (CSO). An unsupervised learning method called the suggested CSO-DCNN employs ReLu by making the inputs and outputs equal. The CSO-DCNN is classifies the fraud activity using real-time and other available dataset that is compared with the existing algorithms. The experimental results shows that the suggested CSO-DCNNobtains98.20%, 99.77%, 95.23% of accuracy for credit card, insurance and Mortgage Data set.