MMFS-CF: A personalized data-driven credit card fraud detection model based on multi-modal multi-objective feature subset selection

An oversampling-integrated deep neural network for imbalanced credit card fraud detection

This study aimed to examine the influence of financial development and financial technology (fintech) on financial inclusion in G20 member countries. The study explored how traditional financial instruments, such as credit cards, debit cards, and ATMs, as well as fintech services, contribute to improving access to formal financial services. Additionally, the study analyzed the impact of economic conditions, particularly inflation, on cashless transactions, as an indicator of financial inclusion. Using a multi-year panel dataset covering G20 economies from 2012 to 2023, the research applied panel-data regression models with country and year fixed effects, controlling for macroeconomic variables such as inflation, market capitalization, and bank branch density. The results revealed that the use of credit cards and electronic payments is positively associated with enhanced financial inclusion. Conversely, the number of ATMs shows a significant negative effect, indicating that the availability of traditional banking infrastructure encourages the continued use of cash, thereby hindering the adoption of cashless transactions. On the other hand, fintech, including e-money accounts, is found to have a highly significant influence on increasing cashless transactions, highlighting the critical role of financial technology in expanding access to formal financial systems. Furthermore, inflation has a significantly negative impact on financial inclusion, as high inflation reduces purchasing power and diminishes trust in formal financial systems. These findings underscore the importance of supporting fintech, reducing reliance on physical banking infrastructure, and maintaining stable macroeconomic policies to promote financial inclusion in G20 member countries. This study provides clear policy implications regarding the necessity of advancing digital financial services to achieve broader financial inclusion.

Credit card-based financial transactions form the backbone of today's digital economy, giving millions of consumers unparalleled convenience in purchases, online payments, and fund transfers. The widespread adoption of credit card-based transactions has concurrently opened avenues to potential fraud-related losses leading to significant losses among individuals, businesses, and financial organizations. Traditional fraud detection systems, which are rule based, give preliminary protection but are not ready to cope with complex and variable fraud patterns that adapt over time. This calls for intelligent fraud detection techniques that apply machine learning to analyze the behaviours of transactions and single out fraudulent activities from the genuine ones. It presents a credit card fraud detection system developed using a Decision Tree classifier, considering interpretability, operational efficiency, and modeling of non-linear decision boundaries. This model examines features in anonymized transactions in order to find unusual patterns that set them apart from typical customer profiles. This roadmap contains extensive data preprocessing, the removal of duplicates, normalization, and class imbalance. Experimental evaluation showed that the Decision Tree classifier yielded reliable performance on fraud transaction detection while maintaining a low false positive rate. This proves that interpretable machine learning models can be embedded into real-world financial systems to enhance security, transparency, and trust in digital transactions.

Financial well-being is increasingly recognized as a structural driver of psychological health, yet empirical evidence in academic backgrounds remains dominated by perception-based measures. This study investigates how financial well-being shape psychological well-being among degree college faculty in Odisha. Using cross-sectional data from 164 faculty members, psychological well-being is measured via the WHO-5 Well-Being Index, while financial well-being is captured through disaggregated ratio-based indicators: EMI burden ratio, savings rate, emergency fund coverage, asset liquidity ratio, and credit card reliance. Ordinary Least Squares estimation with comprehensive diagnostic testing reveals that debt burden exerts a strong and negative effect on psychological well-being, whereas savings capacity, emergency preparedness, and liquidity buffers significantly enhance mental well-being. Credit card reliance shows a negative but comparatively weaker association once structural financial buffers are controlled. Demographic and institutional characteristics lose explanatory power after accounting for financial structure, indicating that everyday financial constraints dominate personal attributes in shaping faculty well-being. By replacing perception-based indices with measurable financial ratios, this study advances the literature on financial well-being and provides rare subnational evidence from India. Keywords: Financial Well-Being; Psychological Well-Being; Debt Burden; Financial Resilience; Higher Education

The development of the cashless society era has driven shifts in financial transaction behavior, particularly among Generation Z, who tend to prefer digital services. However, credit card adoption in this segment remains relatively low, highlighting the need to analyze the factors influencing the intention to use Bank Mega’s Gen-Z credit card. This study examines the effects of security, privacy, and reputation on attitude and intention, with perceived risk and trust serving as mediating variables. A quantitative approach using SEM–PLS was applied to 400 Gen-Z respondents aged 18–27 years residing in Jakarta. The findings reveal that all variables fall into the “very good” category, indicating positive perceptions of security, privacy, reputation, and product reliability. Security, privacy, and reputation significantly enhance trust, while privacy also reduces perceived risk. Conversely, security and reputation do not significantly influence perceived risk. Furthermore, trust has a positive and significant effect on both attitude and intention. Perceived risk influences attitude but does not directly affect intention. Attitude significantly influences intention, suggesting that a positive attitude is a key determinant of Gen-Z’s adoption intention. Overall, the results underscore the importance of strengthening security, privacy, reputation, and risk management in building Gen-Z’s trust. These insights provide strategic implications for Bank Mega in designing more effective marketing approaches tailored to the characteristics of younger consumers.

The advancement of electronic banking has increased the acceptance and use of credit card rendering it as one of the most universally accepted method of payment globally. The incidence of transaction fraud required an effective detection technique to protect customers and financial companies from being trapped by fraudsters. The process of fraud detection, which pertains to the recognition of illicit activities within banking systems, is critical for ensuring financial stability, protecting customer interests, managing institutional reputation, and complying with regulatory requirements. The methodologies encompassing machine learning and deep learning have seen extensive application in addressing issues related to credit card fraud; however, a significant proportion of these methodologies encounter challenges, including erroneous classification and false positives, among other complications. Recent research shows that fraudsters persist in employing novel methodologies in their illicit activities by altering the characteristics or trends in their deceptive practices, thereby rendering fraudulent transactions indistinguishable from genuine ones in an effort to evade detection by current detection mechanisms. To optimize model precision and enhance fraud detection using deep learning feature selection (FS) is of paramount importance. This will alleviate the adverse impacts of noisy, irrelevant and redundant attributes present within the dataset. This research work proposed a new approach that uses Flower Pollination Algorithm (FPA) with Spike Neural Network (SNN) a deep learning technique called FPA-SNN for credit card fraud detection. Four datasets were used to implement the proposed approach, two of the datasets are highly unbalanced with a <1% positive class. To improve classification accuracy and precision we used Synthetic Minority Oversampling Technique (SMOTE) to solve the imbalance problem in the datasets. Realizing that the vast majority of studies in credit card fraud detection uses very few performance metrics for evaluating various machine learning and deep learning algorithms, we utilized multiple evaluation metrics; Accuracy, Precision, Recall, F1-score, the area under curve and the receiver operating characteristic curve (AUC_ROC), and Matthew’s Correlation Coefficient (MCC) to test and evaluate the performance of our proposed model. Our Proposed model performed significantly well with highest MCC greater than 97 percent, as well as AUC-ROC greater than 99.9 percent which shows how robust the model is in feature selection and classification.

The present study examined the association between financial knowledge and credit card management behavior of American adults using a pooled cross-sectional dataset from the National Financial Capability Study 2015, 2018, and 2021. The current study replicated the previous research and improved the methodology using newer datasets. This study constructed a scale for credit card management behavior as a continuous fractional response variable over a defined bounded range. Therefore, fractional regression analysis was used to examine the association between financial knowledge and credit card management behavior. The results suggest that American adults with higher levels of objective and subjective financial knowledge are more likely to engage in positive credit card behavior, and the results are consistent across all income categories. Also, the study treatment for the don’t know (DK) response preserves the construct validity of the objective financial knowledge scale and shows that the association between financial behavior and financial knowledge can be weakened if DK responses are treated incorrectly.

Retraction Note: Optimizing credit card fraud detection with random forests and SMOTE

Detection of credit card fraud has lately been considered a critical task due to the highly imbalanced nature of financial transaction databases. On the other hand, the traditional classification algorithms have been poor at detecting fraudulent activities with an acceptable false-positive rate. Hence, this work contributes to a GMM based approach for fraud detection, which benefits from GMM probabilistic-based classification power for better classification results. The database used in this study is publicly available and was obtained from the Kaggle Credit Card Fraud Detection (CCFD) database. The dataset has 284,807 transactions, and only 0.17% of the cases represent fraud. This paper plans to scale the features, train the GMM technique with different numbers of Gaussian components (i.e., 2, 4, 6, 8, and 10), and evaluate their performance with several evaluation metrics. Compared with other traditional classifiers (logistic regression (92.4%), K-nearest neighbors (93.68%), decision tree (88.16%), and support vector machine (94.21%)), the proposed GMM algorithm obtains the highest accuracy of 94.53%. The proposed method, despite its high accuracy, has limitations under high-dimensional feature dependencies and optimal component selection. From the results obtained over the experimentation process, the GMM proves to be a probable, yet flexible and subservient framework for a complex modelling of probabilities for the detection of fraud.

This paper examines how Unified Payments Interface (UPI) and credit cards are reshaping India’s payment ecosystem and forcing banks to rethink their business models. UPI, launched in 2016, has revolutionized real-time, low-cost transactions and widened financial inclusion, while credit cards continue to dominate in high-value purchases and global acceptance. The study highlights how demonetization, government initiatives, and the pandemic accelerated digital adoption, but also exposed barriers such as limited digital literacy, rural connectivity, and risks of fraud or outages. Comparative analysis shows that while UPI enables microtransactions and empowers small merchants, its zero-MDR regime reduces banks’ revenue streams, unlike credit cards which generate significant interchange and interest income. Emerging innovations; UPI Lite, Credit on UPI, Aadhaar Enabled Payment System, and the Digital Rupee- are explored as potential disruptors and enablers of convergence. The findings suggest coexistence in the short term, with long-term convergence likely to redefine banking, revenue models, and financial inclusion in India.

There are apps for everything, from online banking, social software to shopping. They have become one of the most important tools in our daily lives. This even implies that a mobile device has stored most of personal information, including photos, credit cards, and communications. If an intruder succeeds in hacking into the mobile device, all private properties must suffer from the leakage threats. Undoubtedly, a malware is the commonest tool used by an attacker to compromise a mobile phone. In particular, it is often disguised as a popular application through an obfuscated or packed form. That is the main reason why it is difficult to distinguish a malware from the legal ones. In this article, we have adopted machine learning technique to develop a static analysis mechanism for Android malware detection based on shallow feature and permission correlation (AMD). AMD first analyses the Application Programming Interfaces of the target to detect all possible and hidden privilege threats. It then filters this obfuscation information using permission correlation to eliminate noise and identify meaningful malicious indicators. The proposed approach leverages the correlation patterns between permissions and API calls to distinguish suspicious behaviours from legitimate ones. Thus, AMD can extract all the representative shallow features to achieve the high detection rate. Simulation results have shown that AMD can outperform related works under the datasets of CICAndMal2017 and CICMalDroid2020, which confirms the effectiveness of shallow features and permission correlation.

In data mining and exploration, outliers are specific and infrequent data that require special attention, as they may reveal potentially hazardous information. Detecting outliers can support, e.g., identification fraudulent credit card usage or unauthorized access to transactions, even hacking banking systems, etc. The paper proposes a definition of outlier in terms of fuzzy representations of expert knowledge and its application to detect outliers. The approach proposed has the potential to enhance the performance of outlier detection in various fields, including finance and banking data storage and analysis. By “enhance” we mean that the intention of the new method is to cooperate with known numerical methods, e.g., LOF, rather than supersede or deprecate them. The usefulness of the method is proven via providing new outlying observations for given datasets using input data expressed in an imprecise, linguistic manner.

Real-world data in domains such as finance and fraud detection can be rare, imbalanced, or inaccessible, necessitating synthetic data as a crucial alternative. Gathering and leveraging real-world data in such domains is subject to important challenges such as privacy issues, legality, high cost of annotation, and restricted access due to proprietary ownership. Synthetic data generation in this context offers a meaningful alternative to real data gathering, reducing both privacy and computational costs while allowing for the construction of flexible, scalable datasets. This paper presents a new paradigm for tabular data synthesis through CTGAN (Conditional Tabular GAN) with integration into agentic workflows and retrieval-augmented generation (RAG). The proposed system herein accepts partial data samples and column constraints as inputs from a user-friendly chatbot interface and augment the dataset intelligently through an AI-agent-based generation pipeline. These AI agents aid in the automation of preprocessing, column semantics interpretation, and the enforcement of user-specified constraints specified in natural language, minimizing manual intervention by a considerable margin. The framework further includes ChromaDB to enable semantic retrieval of past relevant datasets. With this semantic memory, the model can improve generation quality, apply schema-level consistency, and update even synthesis of new datasets based on column names or metadata alone. It allows for context-aware, structurally sound, and domain-conformant data generation—without the need to access sensitive or full datasets. The current research utilizes statistical measures like mean, variance, and the Kolmogorov–Smirnov (KS) test to confirm the fidelity of data produced. The approach maintains a mean difference of just 0.16% and a KS statistic of 0.0020, which reflects outstanding statistical consistency with original distributions of data. Preliminary results show significant enhancements in data realism, diversity, and variability without sacrificing domain coherence. The system introduced is particularly well-adapted to financial datasets, such as applications in credit card fraud detection, and offers a scalable, privacy-aware method of synthetic data generation in sensitive or data-scarce environments.

The objective of this study is to analyze the effect of transaction values using cards (Card Payment Instruments/APMK), namely credit cards and debit cards, as well as float funds on the Velocity of Money (VOM) in Indonesia. This study employs a quantitative approach using descriptive methods, utilizing secondary data on credit card transaction values, debit card transaction values, float funds, and VOM over the period 2015–2024. The data sources were obtained from the official websites of the Bank of Indonesia (BI) and the Central Statistics Agency (BPS). The analysis method employed is the Error Correction Model (ECM) to observe short-term and long-term dynamics. The findings of the study suggest that, over an extended period, credit card transaction values (KK) and float funds (DF) exert a positive and significant influence on VOM, while debit card transaction values (KD) prove to be statistically insignificant. In contrast, in the short term, float funds (D(DF)) exert a negative and significant influence on changes in VOM, while changes in credit cards (D(KK)) and debit cards (D(KD)) do not demonstrate a statistically significant effect. Furthermore, the lagged value of credit cards exhibits close to significant at the 10% level, suggesting a gradual adjustment effect on VOM. These findings suggest that policies promoting the optimization of credit cards and the management of float liquidity can encourage money circulation in the long term. However, the impact of debit cards must be carefully considered to avoid directly suppressing the speed of money circulation in the short term. Thus, the evolution of non-cash payment systems must strike a balance between transaction efficiency and its implications for national monetary dynamics.

This study examines how changes in residential permanency during the COVID-19 stay-at-home orders impacted street and cybercrimes in Belo Horizonte, Brazil. Drawing on routine activity theory, we hypothesize that street crimes (theft, auto theft, residential burglary, and robbery) were negatively associated with residential permanency, measured by residential presence from Google Mobility Reports. In contrast, we expect cybercrimes (online fraud) to be positively associated, as increased time at home likely led to greater online activity and exposure to digital victimization. ARIMA time-series models confirmed these predictions for all street crimes except robbery, indicating that offenders adapted to the new circumstances. Surprisingly, cybercrimes were also negatively associated with residential permanency. Ad hoc analysis suggests a potential association between cybercrimes and street crimes. When electronic devices are stolen or private information is accessed (including credit cards, documents, passwords, and other confidential data), perpetrators can utilize these resources to commit further offenses in the digital domain. As a result, the decrease in street crimes may have mitigated the potential catalyzing effect of the orders on cybercrimes, highlighting the need for new theoretical frameworks.

Accurate credit card fraud detection is vital for protecting financial systems and reducing economic losses. Graph neural networks (GNNs) have shown strong potential by capturing complex patterns in transaction networks. However, existing GNN-based approaches exhibit limitations in handling class imbalance, adapting to non-graph transaction data, and capturing the relative importance of features. Therefore, we propose HMOA-GNN, a novel framework for credit card fraud detection designed to handle tabular and highly imbalanced transaction data. First, the density-driven hierarchical hybrid sampling (DEHS) module balances the dataset by generating synthetic fraudulent transactions in dense regions and removing noise. Next, the metric-optimized latent space similarity graph construction (MOLS-GC) module applies metric learning to build graphs that satisfy the homophily assumption. Finally, the Adversarially trained, feature-adaptive GraphSAGE-based model (AdaAdvSAGE) enhances feature aggregation through adversarial learning and adaptive feature selection. Experiments on multiple real-world datasets demonstrate the superior performance of our framework in credit card fraud detection.

Beyond past-due bills: The varieties of medical debt used to finance healthcare.

The purpose of this project is to examine the opportunities and barriers of the E-payment system in Bangladesh. Now a day’s world changed to the digital world and also Bangladesh tries to step towards Vision 2041. E-payment is a method in which a person can make online payments for their purchase of goods and services without the physical transfer of cash and cheques, irrespective of location and time. In the present scenario, this study aimed to identify “Opportunities and Barriers of E-payment in Bangladesh” differ on different generations using the quantitative method and offer some solutions to improve the e-payment system. Qualitative research is the main methodology in this project with face-to-face interviews and semi-structured questionnaire methods. Data of survey was collected from a total of 50 participants who include undergraduate and postgraduate college students by conducting a series of semi-structured face-to-face interviews and questionnaire methods. Out of 50 respondents were found to be fully completed and satisfactory for analysis. The study also found that people who are using and are likely to use e- payment preferred debit cards, e-wallet as the first mode of payment than credit cards. This study suggests that the government of Bangladesh should set up a Regulatory Commission on an urgent basis to strengthen the e-payment system.

This study discusses the use of classical and quantum machine learning models to detect fraudulent bank transactions. Random Forest model was tested on credit card fraud detection data set and scored large percentage 99.95, AUC-ROC score/ROC is 1.0 and F1 scores are high. The most influential predictors were identified to be key features including the amount of transaction, periods between transactions, and location. In order to avoid the problem of class imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) was utilized, which enhanced the work of the model. Another promising study of quantum hardware scalability limits, but with multiple serious limitations, was the Quantum Support Vector Classifier (QSVC), which faces difficulty in qubit coherence and scalability challenges. These limitations did not allow the model to effectively process large data sets to better accommodate real world applications. Nevertheless, quantum models have the potential to improve the fraud detection system with developing quantum technology. This study brings out the usefulness of Random Forest in detecting fraud cases and outlines the opportunities of quantum models in the future, recommending future research, such as quantum-classical hybrid models, and the enhancement of quantum computers to meet real-time needs.