International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 05 | May 2024
p-ISSN: 2395-0072
www.irjet.net
A Survey on credit Card Fraud Detection using machine and Deep Learning Megha Nayak1, Prof. Satendra Sonare2 1Reseacrh Scholar, Department of CSE, Gyan Ganga Institute of Technology and Sciences, Jabalpur, M.P. 2Professor, Departmet of CSE, Gyan Ganga Institute of Technology and Sciences, Jabalpur, M.P.
---------------------------------------------------------------------***--------------------------------------------------------------------identify this type of strain is to analyze each account's payment pattern and identify errors associated with Recent developments in e-commerce and "normal" transaction design [6]. Unauthorized transaction telecommunications have increased the use of credit cards by someone other than the owner of the credit card or for both online and everyday transactions. However, cases of account. A stolen, lost or fraudulent credit card may result in credit card fraud are on the rise, causing financial cancellation. As online shopping has increased, cardless institutions to incur huge losses every year. Establishing extortion or the use of credit cards in e-commerce has effective fraud detection systems is critical to reducing these become more common. The spread of frauds such as ccf losses, but is difficult due to the instability of most credit resulting from the proliferation of electronic banking and, in card information. Additionally, credit card fraud using some cases, online payments leads to losses of billions of traditional machine learning algorithms is ineffective as their dollars each year. In the era of computer programming, ccf structure consists of a static map from input vectors to detection has become one of the most important goals. As a vectors. For this reason, customers who use credit cards business owner, it doesn't matter if long-term transportation cannot transfer their purchases. A clustering-based is moving towards a money-free culture. As a result, transformation using a neural network as the learner is traditional payment strategies may not be available in the adopted. The performance of the plan has been confirmed future and may not be suitable for business continuity. using evidence available in the credit card world. Customers do not need to come into the store with cash in their pockets. They now charge bills and credit cards. Keywords: Detection of fraud; tracking of fraud; Fraud Therefore, companies need to adapt their environments to transaction understanding, Neural Network, Adaptive enable them to accept a variety of payments. This situation is Learning. expected to continue for a long time [7]. CCF was the top source of theft recorded this year, after government 1. INTRODUCTION documents and financial support [9]. In 2020, 365,597 attempts were made to spend money using an unused credit Electronic commerce in general has increased over time due card [10]. The number of theft-related conduct complaints to the popularity of e-commerce such as online stores such increased 113% from 2019 to 2020; this includes a reported as Amazon, eBay and Alibaba. Credit/debit cards are widely 44.6% increase in credit card identity theft [11]. Payment used in electronic transactions. Recently, card-not-present card theft cost the global economy $24.26 billion last year. transactions[1] have become more common in the credit Elaborate credit card skimming incidents accounted for card industry, ultimately through online payment systems 38.6% of incidents in 2018, making these states the least such as paypal and alipay. The display value of global eprotected from credit card theft. Therefore, examining the commerce is expected to reach $24 trillion by 2019 [2]. But use of money should be the priority of the use of inheritance. at the same time, fraud occurs, which affects the economy. A The program must identify fake and non-fake swaps and use study of more than 160 companies found that the number of this information to decide whether a future swap is fake or online frauds each year is 12 times greater than offline not. This problem includes major problems such as the frauds [4]. Since there is no need for a physical card in the eoperating speed of the system, attracting attention, and commerce environment and almost all of the information on prioritization. Machine learning is a field of visual science the card is sufficient for transactions, fraudsters need the that uses computers to make predictions based on early data information in this letter to gain power. For example, after patterns. Machine learning models have been used in many scammers obtain card account numbers and passwords from studies to solve many problems. Deep learning (dl) honest cardholders, they use them to purchase certain applications include computer network organizations, products. Fraudsters often obtain card information through anomaly detection, financial management, protection, a variety of means: hijacking attempts to send letters portable cellular systems, medical ransomware discovery, containing the latest releases, copy-pasting of card recovery and malware discovery, video surveillance information by copiers, phishing (clone websites), or from discovery, zone tracking, android malware discovery, home unknown credit card companies. 5]. Due to the complexity of robotization and heart disease prediction. . For data the environment and population base, it is inevitable that all classification, inverse vector machines (svm) can be a legitimate card holders will be robbed. The best way to training method. It has been used in many fields, including
Abstract –
© 2024, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 830