DeLClustE: Protecting Users from Credit-Card Fraud Transaction via the Deep-Learning Cluster Ensemble

Abstract

Fraud is the unlawful acquisition of valuable assets gained via intended misrepresentation. It is a crime committed by either an internal/external user, and associated with acts of theft, embezzlement, and larceny. The proliferation of credit cards to aid financial inclusiveness has its usefulness alongside it attracting malicious attacks for gains. Attempts to classify fraudulent credit card transactions have yielded formal taxonomies as these attacks seek to evade detection. We propose a deep learning ensemble via a profile hidden Markov model with a deep neural network, which is poised to effectively classify credit-card fraud with a high degree of accuracy, reduce errors, and timely fashion. The result shows the ensemble effectively classified benign transactions with a precision of 97 percent. Thus, we posit a new scheme that is more logical, intuitive, reusable, exhaustive, and robust in classifying such fraudulent transactions based on the attack source, cause(s), and attack time gap.

Enhancing the Random Forest Model via Synthetic Minority Oversampling Technique for Credit-Card Fraud Detection

“…With a focus on feature selection -their study achieved a prediction accuracy of 99.2% with reduced training time that did not compromise model performance. Research [46] addressed the challenges in [27] on how fraud acts are masked, examined detection procedures, and analyzed the many motivations for adversaries to exploit fraud actions, threats, and network breaches. They proposed a hybrid modular ensemble for credit card fraud detection, which achieved a prediction accuracy of 99.6% to classify benign from genuine transactions effectively.…”

“…These they learn through features classification either from the normal behavior cum signature in transactions or the quick detection of unusual activity in the transaction pattern indicative of a fraudulent profile. A variety of such machine learning (ML) models that have been successfully used or implemented include Logistic Regression [22]- [24], Deep Learning [25]- [27], Bayesian model [28], Naive Bayes [29], Support Vector Machine [30], [31], K-Nearest Neighbors [32], Random Forest [33], [34], and other models [35], [36] that have been effectively used to detect credit card fraud. Many of these have drawbacks with their flexibility in feature selection, importance, and accuracy.…”

Handling Transactional Data Features via Associative Rule Mining for Mobile Online Shopping Platforms

“…It bolsters the reasoned action through a planned set of behaviors [78] as thus: (a) that a consumer absorbs advertised information and knowledge as presented by the vendor [79], (b) that a consumer may process the retrieved knowledge about a product, and also can leverage on previous experience to compare between the observed versus expected outcome [80], and (c) that a consumer decides either to accept or reject the purchase of an item [81] yielding a choice or decision reached from balanced insight through mental synthesis. Thus, with data input as its greatest prize [82] the product manufacturers must equip managers with adequate knowledge in place of the product line that will eventually drive consumers to keep buying the item; And in turnthis will shore up and push up sales volume of the product [83]. This theory unveils the underlying feats that may cause purchase shift in the consumer behavior [84] such that where and if a consumer is not adequately informed, s(he) can reject the purchase of an item as means to normalize with the online data cum knowledge available [85].…”

FePARM: The Frequency-Patterned Associative Rule Mining Framework on Consumer Purchasing-Pattern for Online Shops

“…The framework must flexibly exploit new itemset generated on the fly, and robustly cum easily reuse such a dataset with little (or no) modification to the system. This solution seeks to ensure: (a) distance between itemsets, is retained, (b) sales volume of two/more commonly purchased itemset, are positively correlated, (c) if/when itemsets sales volumes are observed at different timestamps, it generates a trend that must be similar in their upward/downward temporal patterns, with both trends assigned to same cluster -so that each cluster is a set of items analogous to an itemset in association rule mining [31]- [33].…”