期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2019
卷号：7
期号：5
页码：3173-3177
DOI：10.15680/IJIRCCE.2019. 0705095
出版社：S&S Publications
摘要：Nowadays, the Internet is an important factor of our life. Dueto the wide use of the internet, the status
of online shoppingis varies day by day. The Credit Card is the easiest method for online shopping and paying bills.
Therefore, Credit Card becomes popular and appropriate approach for online money transaction and it isgrowing very
quickly. In this paper, machine learning algorithms are utilized for the detection of credit card fraud. Firstly, common
type of models is used. After that, hybrid methods which can use to Ada Boost and majority voting methods are
activated. Ada Boost method is able to develop the individual results from different algorithms. To estimate the model
efficiency, an openly accessible credit card data set is used. After that, a real-world credit card dataset from a financial
organization is evaluated. In addition, noise is added to the examples of data to further evaluate the toughness of the
algorithms. In this paper, to classify the most important variables that can guide to superior accuracy in credit card
fraudulent transaction detection technique. Additionally, we explain the performance of different supervised machine
learning algorithms that are existed in literature against the good classifier that it executed in this paper. The final
results of this system have positively identified that the majority of voting method obtains better quality, accuracy
ratios in catching fraud cases in credit cards for identification of actual credit card transaction data.
