Credit Card Transactions Research Articles

An Ensemble-Based Credit Card Fraud Detection Algorithm Using an Efficient Voting Strategy

Abstract The existence of fraud in credit card transactions causes many financial losses leading to customers’ loss of trust. Fraud detection methods based on machine learning techniques prevent such losses. Despite the literature on fraud detection, there is a lack of algorithms that detect fraud with acceptable performance in the credit card fraud detection field. Therefore, this study proposed a comprehensive ensemble-based method using an efficient weighted voting strategy for credit card fraud detection that can address the previous algorithms’ weaknesses. First, since the dataset is imbalanced, the proposed method balanced the dataset by stratifying it into three different proportions of normal and fraudulent transactions (1 to 1, 1 to 4 and 1 to 9 ratios). The features in each dataset are ranked by four feature-ranking methods, and the Random Forest classifier is applied to each of them for selecting the essential features. Afterward, using the seven base classifiers and the chosen features, 12 ensembles have been developed. Besides, a weighted voting strategy is proposed, and the fraudulent transactions are detected through voting based on the base classifiers’ and ensembles’ weights, which are calculated by their performance. The computational results indicated that the suggested Eclf10 is the best ensemble and its Logistic Regression classifier also has the best performance among other base classifiers. The Eclf10 leads to 99.97% accuracy, 87.78% precision, 97.70% recall, 92.21% F1-score and 95.634% F2-score, which has a superiority over the previous ensemble-based methods (e.g. majority voting ensemble, stacking classifier, Adaboost, Gradient Boosting).

Non-Intrusive Assessment of COVID-19 Lockdown Follow-Up and Impact Using Credit Card Information: Case Study in Chile.

In this paper, we propose and validate with data extracted from the city of Santiago, capital of Chile, a methodology to assess the actual impact of lockdown measures based on the anonymized and geolocated data from credit card transactions. Using unsupervised Latent Dirichlet Allocation (LDA) semantic topic discovery, we identify temporal patterns in the use of credit cards that allow us to quantitatively assess the changes in the behavior of the people under the lockdown measures because of the COVID-19 pandemic. An unsupervised latent topic analysis uncovers the main patterns of credit card transaction activity that explain the behavior of the inhabitants of Santiago City. The approach is non-intrusive because it does not require the collaboration of people for providing the anonymous data. It does not interfere with the actual behavior of the people in the city; hence, it does not introduce any bias. We identify a strong downturn of the economic activity as measured by credit card transactions (down to 70%), and thus of the economic activity, in city sections (communes) that were subjected to lockdown versus communes without lockdown. This change in behavior is confirmed by independent data from mobile phone connectivity. The reduction of activity emerges before the actual lockdowns were enforced, suggesting that the population was spontaneously implementing the required measures for slowing virus propagation.

CREDIT CARD FRAUD DETECTION AND ANALYSIS USING MACHINE LEARNING ALGORITHMS

Increase in the number of customers since last decade for credit card usage. The customer’s without prior approval many frauds have been detected. The unethical use of credit cards by hackers or credit cards users unwilling to pay back the amount are known as the major credit card frauds. The credit card transactions should be detected earlier by statistical methods if any fraud has been detected. The various algorithms in the machine learning are used to analyze the patterns and frauds. This helps the bank to eliminate the frauds by declining suspected transactions.

WHY FINANCE & BANKING SECTOR NEED BIG DATA?

Have you ever thought of the amount of data you create each day? Every message you send, every credit card transaction, even every web page you open… They all collect a total of 2.5 quintillion bytes of data produced daily by the global population. This offers endless opportunities to leverage this data for the most forward-thinking businesses in many areas, and the banking industry is no exception While digital banking is used by almost half of the world's adult population, financial institutions have enough data at their disposal to rethink the way they work, becoming more efficient, more customer-focused, and ultimately more profitable.

Thy Neighbor’s Misfortune: Peer Effect on Consumption

Using a large, representative sample of credit and debit card transactions in Singapore, this paper studies the consumption response of individuals whose same-building neighbors experienced personal bankruptcy. The unique bankruptcy rules in Singapore suggest liquidity shocks drive personal bankruptcy decisions, leading to a substantial drop in consumption for the bankrupt. Peers’ monthly card consumption decreases by 3.4 percent over the 1-year postbankruptcy period. There exists no consumption decrease among individuals in immediately adjacent buildings nor for consumers with diminished postevent social ties with the bankrupt. The findings imply a significant social multiplier effect of 2.8 times the original consumption shock. (JEL E21, G51, K35)

Machine Learning based Fraud Analysis and Detection System

The spectacular surge in the proportion of credit card transactions, web based purchases, has led to a surge in fraudulent activities recently. For any business establishment, credit card security is a major concern. In this respect, credit card fraud is hard to identify. Thus it became imperative to implement effectual fraud detection systems for all credit card issuing banks to mitigate their losses. Betrayed transactions with real transactions in actuality are often dispersed and simple methods of matching are not enough to detect them accurately. The paper proposes an algorithm based on Machine Learning credit card fraud detection to solve the issue of a fraudulent transaction. This framework nominally increases the probability of card fraud by exponential activity. The results show that the accuracy of Random Forest, Support Vector Machine and KNN classifiers achieves respectively 94.84%, 89.46%. Random Forest could even predict new fraud cases very quickly.

Accurate fraud detection in credit card transactions using hybrid heuristic and meta-heuristic algorithms

Credit card based online transactions, payment methods developed rapidly, Credit card fraud also additionally expanded simultaneously. The particular crime in banking system is credit card fraud. The main objective of proposed AFD-HHM technique is, to detect the fraudulent transaction with high accurate rate and solve the user side problems. Some most dangerous credit card crimes are credit card application fraud, phishing scams, credit card skimming, online sales fraud, Credit Card Imprints, which leads to following user problems such as class imbalance problem, verification latency and concept drift. This paper focus to detect the fraud transactions with higher efficeiency rate solves the user side problems, high computational complexity problems. Initially, we propose Improved Egyptian Vulture Optimization (IEVO) algorithm (meta-heuristic) to grouping all cardholders into different groups based on the transaction behaviours. Second, we can aggregate each group transactions and compute the optimal attributes for each cardholder by using of trust rule strategy. Third, we propose Differential Evolution based Neural Network (DENN) classifier for differentiate transaction as normal or fraud. The result of proposed approach gives better accuracy with higher rate which is comparing with existing methods.

Empty Summer: International Tourist Behavior in Spain during COVID-19

COVID-19 has led to the biggest tourism crisis in modern Spanish history. Although its borders were opened in June 2020, incoming international tourism to Spain was very low. Tourist behavior, however, has been very diverse between types of tourism. Data from credit card transactions showing tourist behavior in 33 destinations representative of various types of tourism were analyzed. Results show that differences in tourist behavior between destinations can be explained by risk perception and the proximity of the destination to home, which favors natural areas and charming coastal cities. This study contributes to the academic literature by studying the effective behavior of international tourists in the time of COVID-19.

Investigating Cybercrimes In Malaysia: The Importance Of ECommerce Security To Eliminate Fraud And Security

The e-commerce marketplace in Malaysia is showing an advantageous trend, with gross value added contributed 6. 3% to gross domestic product (GDP) in 2017.Online purchasing has attained a lot of important factors in the current marketing. But, because of the number of growth of trouble makers or scams had increased, such activities will create a fear in the mind of customers and as well can influence the attitude or mind set of consumers toward online purchase. Therefore, the issue in this area of this survey is customer’s dissatisfaction and attitudes towards online purchasing.Apart from that, customers also concern about the security level whenever they order products online. The security refers to online transaction safety. The previous researchers suggest that the lack of security on the website will make the customers to have doubt or serious concern about their credit card transactions and privacy of personal information to avoid cybercrimes.In spite of comfort and advantages of online shopping, e-commerce poses a few demanding situations and concerns amongst consumers on security and privateness problems. Thus, this phase intends to identify degree of situation in regards to these issues. In widespread, slightly extra than half of on-line buyers claimed that they are problem on safety and privacy at the same time as buying online. The regression analysis indicates fraud and security was found to have significant at 5% significant level (sig, t = 0.004). This could explained that customers are more likely to take serious on fraud and security of the products they purchased.

An Efficient Techniques for Fraudulent detection in Credit Card Dataset: A Comprehensive study

Now a day, credit card transaction is one the famous mode for financial transaction. Increasing trends of financial transactions through credit cards also invite fraud activities that involve the loss of billions of dollars globally. It is also been observed that fraudulent transactions have increased by 35% from 2018. A huge amount of transaction data is available to analyze the fraud detection activities that require analysis of behavior/abnormalities in the transaction dataset to detect and ignore the undesirable action of the suspected person. The proposed paper lists a compressive summary of various techniques for the classification of fraud transactions from the various datasets to alert the user for such transactions.

Credit card payments: do cultural values matter? Evidence from the European Union

PurposeThis study analyzes how cultural and social values shape specific attitudes toward credit cards and indebtedness and consumption behavior.Design/methodology/approachThe study uses a panel dataset for a selection of European Union countries from 2003 to 2016. The relation between credit card use and social and cultural attitudes is constructed by controlling for past habits in payment behavior and cross-substitution with alternative payment instruments by employing a dynamic panel data analysis based on the system Generalized Method of Moments (GMM) estimator.FindingsThe total value of credit card payments positively correlated with values emphasizing risk-taking attitudes. When analyzing the propensity of using these instruments for larger purchases, the level of trust is the most relevant predictor. However, the results seemed region-specific with some variables correlating consumption behavior with credit card usage depending on the political and the economic background of the country. Moreover, risk-taking attitudes prevail when they are related to the extent to which countries rely on cash as a preferred payment instrument. Also, credit card usage is mainly explained by past habits and the economic context.Originality/valueThe model expands on previous credit card transaction research by including an additional set of cultural values able to account for the complex nature of payment instruments and their effects on indebtedness and consumption behavior.

Benefits of Mobile Contact Tracing on COVID-19: Tracing Capacity Perspectives.

For effectively suppressing COVID-19's spread, contact tracing has been widely used to identify, isolate, and follow-up with those who have come in close contact with an infected person (or “close contacts”). Traditionally, contact tracers in local health offices interview an infected person to identify visited places (or hotspots) and then check any close contacts. For the accurate recall of travel history, several countries including South Korea corroborate multiple data sources, such as cell location or credit card transactions (1). Beside this traditional approach, various mobile apps were introduced to help improve travel history tracking including automated GPS tracking (e.g., Israel's HaMagen) and manual place QR-code scanning (e.g., New Zealand's NZ COVID Tracer and Korea's KI Pass). Alternatively, mobile apps maintain individuals' “encounter history” (instead of place visit history) by leveraging peer-to-peer wireless beaconing (i.e., self-announcing its presence to nearby devices) with Bluetooth Low Energy (BLE) in smartphones, such as Google-Apple's Exposure Notification and Singapore's BlueTrace and TraceTogether. This encounter history can be used later to judge whether a user had a risky encounter with an infected person. We argue that traditional manual contact tracing can be greatly improved by leveraging the wisdom of crowds. Local community members install mobile apps to self-collect “breadcrumbs” for contact tracing, such as GPS traces, place QR-codes, and wireless encounter histories, which can offer near real-time assessment (2). However, there is a systematic lack of adoption of mobile apps in many countries, and success stories of mobile apps are scarce. This is partly because local health authorities are unsure about the benefits of mobile apps, mostly worrying that a lack of app adoption fails to achieve digital herd immunity (3). We emphasize the importance of adopting contact tracing apps via “contact tracing capacity modeling.” Our model clearly demonstrates that the adoption of contact tracing model apps can potentially increase contact tracing capacity. Furthermore, this capacity increment thanks to crowd participation will greatly help local authorities to better handle confirmed cases in the early or late phases of the pandemics. In the following, we present our approach of contact tracing capacity modeling and its benefits for contact tracing. We then discuss practical challenges related to application adoptions by citizens and health authorities as well as technical implementation issues.

Explainable Credit Card Fraud Detection with Image Conversion

The increase in the volume and velocity of credit card transactions causes class imbalance and concept deviation problems in data sets where credit card fraud is detected. These problems make it very difficult for traditional approaches to produce robust detection models. In this study, a different perspective has been developed for this problem and a novel approach named Fraud Detection with Image Conversion (FDIC) is proposed. FDIC handles credit card transactions as time series and transforms them into images. These images, which comprise temporal correlations and bilateral relationships of features, are classified by a convolutional neural network architecture as fraudulent or legitimate. When the obtained results are compared with the related studies, FDIC has the best F1-score and recall values, which are 85.49% and 80.35%, respectively. Since the images created during the FDIC process are difficult to interpret, a new explainable artificial intelligence approach is also presented. In this way, feature relationships that have a dominant effect on fraud detection are revealed.

The COVID-19 Shock and Consumer Credit: Evidence from Credit Card Data

We use credit card data from the Federal Reserve Board's FR Y-14M reports to study the impact of the COVID-19 shock on the use and availability of consumer credit across borrower types from March through August 2020. We document an initial sharp decrease in credit card transactions and outstanding balances in March and April. While spending starts to recover by May, especially for risky borrowers, balances remain depressed overall. We find a strong negative impact of local pandemic severity on credit use, which becomes smaller over time, consistent with pandemic fatigue. Restrictive public health interventions also negatively affect credit use, but the pandemic itself is the main driver. We further document a large reduction in credit card originations, especially to risky borrowers. Consistent with a tightening of credit supply and a flight-to-safety response of banks, we find an increase in interest rates of newly issued credit cards to less creditworthy borrowers.

신용카드 사용내역을 통한 해외여행 예측 모형의 개발

본 연구에서는 신용카드 사용내역을 통해 고객의 라이프 이벤트 중 해외여행을 예측하는 모형을 만들고, 예측력이 높은 모델을 구성하는데 있어서 텍스트 분석 방법인 FastText 방법론과 새롭게 제안하는 adaptive weighted sum of term scoring (AWST score) 방법을 기존의 머신러닝 모델과 함께 비교한다. 소비 양상을 통해 소비주체의 삶이 고스란히 나타나는 신용카드 내역 데이터는 각 고객의 라이프 이벤트들을 예측하는 데에 중요한 단서가 되며, 이를 통해 만들어진 예측 모형은 고객이 위치한 현재 삶의 단계나 상황을 더욱 풍부하게 이해하도록 도움을 준다. 본 연구에서는 다양한 라이프 이벤트 다운데, ‘해외여행’을 계획하는 고객을 예측하는 내용을 소개한다. 기존의 머신러닝 기법을 적용한 예측모형 및 새롭게 제안하는 방법의 예측모형의 성능을 정확도, 정밀도, 재현율, 곡선하면적, F1 점수, 양성우도비의 지표를 통하여 비교하고, 개별 모형들을 앙상블 방법으로 결합하여 최종적인 예측 모형을 개발하는 것을 목표로 한다.

The Good, the Bad and the Ugly of Islamic Credit Cards

The Malaysian Islamic financial services have developed and thrived in the competitive domestic and global financial market especially in the last three decades. The Government has provided the industry conducive enabling environment to catalyse the industry growth and development. Islamic finance has gained prominence and been identified as the growth area in the nation’s financial sector. The Bank Negara Malaysia Annual Report 2020 published on 3 April 2020 amongst others highlights that with an advanced regulatory framework already in place, Islamic finance is poised to play a more prominent role in the coming period, particularly in its potential to apply shariah principles to expand social finance and address market gaps in innovative ways. With a range of innovative shariah compliant products and services, the halal financial services have gained market acceptance from both Muslims and non-Muslims alike, albeit with different levels of acceptance according to the products. Specifically, to a certain degree it appears that the market has been more receptive of other Islamic financial services such as loans for various purposes including purchases of securities, properties, vehicles; working capital and even personal loans than the credit cards. According to the Bank Negara Monthly Statistical Bulletin: Monthly Highlights and Statistics in September 2020, as of September 2020, Islamic banks had disbursed 35.7% or RM651.4 billion of the total loans in the banking system in September 2020. The Islamic credit cards however have a lower market share of 10.2% or RM3.7 billion of total credit card transactions. The slower growth pace, lower market share and performance of Islamic credit cards vis–a-vis conventional credit cards, and also compared with other Islamic financial products merit further scrutiny and analysis to help better understand the issues pertinent the growth of the Islamic credit cards. Issues such as why other Islamic loans and hire purchase products have performed better; why the slower growth of Islamic credit cards; customers and market expectation of Islamic credit cards; Islamic credit card product development and marketing strategies need to be carefully examined in order to overcome Islamic credit card growth conundrum.

Credit card fraud detection using artificial neural network

Frauds in credit card transactions are common today as most of us are using the credit card payment methods more frequently. This is due to the advancement of Technology and increase in online transaction resulting in frauds causing huge financial loss. Therefore, there is need for effective methods to reduce the loss. In addition, fraudsters find ways to steal the credit card information of the user by sending fake SMS and calls, also through masquerading attack, phishing attack and so on. This paper aims in using the multiple algorithms of Machine learning such as support vector machine (SVM), k-nearest neighbor (Knn) and artificial neural network (ANN) in predicting the occurrence of the fraud. Further, we conduct a differentiation of the accomplished supervised machine learning and deep learning techniques to differentiate between fraud and non-fraud transactions.

Spectral-Cluster Solution For Credit-Card Fraud Detection Using A Genetic Algorithm Trained Modular Deep Learning Neural Network

Adversaries achieved such intrusion via carefully crafted attacks of large magnitude that seek to wreak havoc on network infrastructures with a focus on personal gains and rewards. Study proposes a spectral-clustering hybrid of genetic algorithm trained modular neural network to detect fraud in credit card transactions. The hybrid ensemble seeks to equip credit-card users with a system and algorithm whose knowledge will altruistically detect fraud on credit cards. Results show that the hybrid model effectively differentiates between benign and genuine credit card transactions with a model accuracy of 74%.

Semi-Supervised Classification on Credit Card Fraud Detection using AutoEncoders

The use of credit cards for online purchases has increased dramatically and led to an explosion in credit card fraud. Credit card companies need to be able to identify fraudulent credit card transactions so that customers are not charged for items they do not buy. In this study, we will use semi-supervised learning and combine it with AutoEncoders to identify fraudulent credit card transactions. In this paper, we will implement the use of T-SNE to visualize fraud and non-fraud transactions, then improve the visualization using autoencoders. Classification report proved that it is possible to achieve very acceptable precision using semi-supervised classification to detect credit card fraud.

A Frequent Pattern Conjunction Heuristic for Rule Generation in Data Streams

This paper introduces a new and expressive algorithm for inducing descriptive rule-sets from streaming data in real-time in order to describe frequent patterns explicitly encoded in the stream. Data Stream Mining (DSM) is concerned with the automatic analysis of data streams in real-time. Rapid flows of data challenge the state-of-the art processing and communication infrastructure, hence the motivation for research and innovation into real-time algorithms that analyse data streams on-the-fly and can automatically adapt to concept drifts. To date, DSM techniques have largely focused on predictive data mining applications that aim to forecast the value of a particular target feature of unseen data instances, answering questions such as whether a credit card transaction is fraudulent or not. A real-time, expressive and descriptive Data Mining technique for streaming data has not been previously established as part of the DSM toolkit. This has motivated the work reported in this paper, which has resulted in developing and validating a Generalised Rule Induction (GRI) tool, thus producing expressive rules as explanations that can be easily understood by human analysts. The expressiveness of decision models in data streams serves the objectives of transparency, underpinning the vision of ‘explainable AI’ and yet is an area of research that has attracted less attention despite being of high practical importance. The algorithm introduced and described in this paper is termed Fast Generalised Rule Induction (FGRI). FGRI is able to induce descriptive rules incrementally for raw data from both categorical and numerical features. FGRI is able to adapt rule-sets to changes of the pattern encoded in the data stream (concept drift) on the fly as new data arrives and can thus be applied continuously in real-time. The paper also provides a theoretical, qualitative and empirical evaluation of FGRI.