Earticle

The anonymity of Bitcoin transaction has let Bitcoin be used as a medium of illicit activities in the dark-net marketplace that are related to crimes such as selling drugs, weapon, etc. Detecting illicit Bitcoin transaction has been drawing attention from government agencies and academia, since FBI’s investigation from 2011 to 2013, when FBI seized a marketplace, SilkRoad, which trades illicit goods and services only with Bitcoin. However, little research has been conducted to detect illicit Bitcoin transactions. In this paper, we applied data mining approach to detect illicit Bitcoin transaction using a dataset which consists of illicit Bitcoin transaction data released by FBI and legal Bitcoin transaction data. We built several classification models such as RandomForest, Decision Tree(C5.0), and SVM. 10-fold cross-validation reveals that RandomForest outperforms the other two. It is expected that we can reduce the investigation time and cost to detect illicit Bitcoin transactions.

The rise of cryptocurrencies has also led to an increase in fraudulent activities, posing challenges in fraud detection on decentralized platforms like Ethereum. This issue is particularly pronounced in decentralized environments like Ethereum, where new transaction patterns continue to emerge. In this dynamic and changing environment, it is important to address the problem of concept drift, which refers to the continuous evolution of data patterns. To address the dynamic nature of these fraud patterns, we propose an automated hyperparameter optimization (HPO) approach using Proximal Policy Optimization (PPO). Unlike traditional HPO methods, PPO efficiently navigates the complex hyperparameter space, adapting to evolving fraud schemes with minimal human intervention. Our method enhances the adaptability and robustness of fraud detection models, effectively improving detection accuracy. Experimental results demonstrate that PPO outperforms existing HPO techniques, offering a more flexible and powerful tool for maintaining the performance of fraud detection systems in the rapidly changing cryptocurrency landscape.

비대면 금융거래의 확산으로 이상 거래 탐지와 서명 위·변조 검출의 중요성이 높아지고 있다. 기존 기술은 단일모달 데이터에 의존해 서명 행위의 동적 특징을 충분히 반영하지 못하였다. 본 연구는 거래 데이터, 서명 이미지, 필기 행위 시계열을 통합 분석하는 멀티모달 인공지능 모델(MAIFDM) 을 제안한다. MAIFDM는 시공간 주의학습, 문맥 임베딩 학습, 시계열 연관성 불일치 학습을 결합하여, 세 모듈의 특징을 융합한 뒤 마할라노비스 거리와 Adaptive 동적 임계치를 통해 이상 여부를 판정한다. 실험 결과, MAIFDM은 기존 모델 대비 F1-score 0.907, AUC 0.942로 우수한 성능을 보임으로써, 멀티모달 데이터 학습과 사기거래 탐지에 효과적인 모델임을 증명하였다.

The importance of detecting abnormal transactions and detecting signature forgery or alteration is increasing due to the proliferation of non-face-to-face financial transactions. Existing technology has not sufficiently reflected the dynamic characteristics of signature behavior by relying on single modal data. This study proposes a multimodal artificial intelligence model (MAIFDM) that integrates and analyzes transaction data, signature images, and handwriting behavior time series. MAIFDM combines time-space attention learning, context embedding learning, and time-series correlation mismatch learning to fuse the features of the three modules and then determines whether there is an abnormality through the Mahalanobis distance and adaptive dynamic threshold. As a result of the experiment, MAIFDM showed superior performance with F1-score 0.907 and AUC 0.942 compared to the existing model, proving that it is an effective model for multimodal data learning and fraudulent transaction detection.

The past has focused on domestic transactions from the point of accident prevention on the basis of security. Of course, it is a problem that its security measures had been focused on the user terminal stage. By supplementing it now, securities business circles are expected to announce "None Face-To-Face Account Opening". Therefore, they have been introducing and advancing Fraud Detection System (FDS) that integrates log data from HTS (Home Trading system) and security system; models trading patterns; and detects and analyzes abnormal financial transactions based on the scenarios. Consider-ing such characteristics of securities business as ordering and transferring stocks or shares, the present study proposes trading patterns specialized in the business and detec-tion plan (decision making) after examine four major functions, patterns and detection of FDS.

This study proposes an AI based on abnormal behavior detection method to identify abnormal transactions that may occur in blockchain based data transactions. For the study, This study utilized the Hyperledger Fabric platform and automatic transactions of smart contracts. The traded data is recorded on the blockchain through smart contracts, and sensitive metadata is stored in off chain storage. As an AI model, XGBoost, which has high interpretability, was used, and abnormal transactions are detected by analyzing user transaction logs through the model, and the trust score is adjusted in real time through smart contracts. The experimental results showed that the performance indicators were improved by about 2~ 5% compared to previous similar studies, and this study can operate a trust based transaction automation system.

Despite all attempts to prevent fraud, it continues to be a major threat to industry and government. Traditionally, organizations have focused on fraud prevention rather than detection, to combat fraud. In this paper we present a role mining inspired approach to represent user behaviour in Enterprise Resource Planning (ERP) systems, primarily aimed at detecting opportunities to commit fraud or potentially suspicious activities. We have adapted an approach which uses set theory to create transaction profiles based on analysis of user activity records. Based on these transaction profiles, we propose a set of (1) anomaly types to detect potentially suspicious user behaviour, and (2) scenarios to identify inadequate segregation of duties in an ERP environment. In addition, we present two algorithms to construct a directed acyclic graph to represent relationships between transaction profiles. Experiments were conducted using a real dataset obtained from a teaching environment and a demonstration dataset, both using SAP R/3, presently the predominant ERP system. The results of this empirical research demonstrate the effectiveness of the proposed approach.

정보기술의 빠른 진화, 빅데이터의 등장, 분석기법의 고도화 등으로 인해 다량의 데이터로부터 의미있는 정보를 추출하는 데이터마이닝을 다양한 영역에 활용하고자 하는 시도들이 활발히 진행되고 있다. 그 중의 한 분야가 농산물 유통영역인데, 농산물에 대한 지속적인 수요 증가와 전자경매의 활성화 등으로 수도권 농산물 도매시장에서만도 연간 수천만건 이상의 거래가 이루어 진다. 그러나 급속한 거래량 증가와 더불어 과거로부터 관행적으로 이루어지고 있는 부정거래도 함께 증가하고 있는데 거래참가자들 사이의 결탁에 의해 발생하는 농산물 도매시장의 부정거래는 점차 지능화되는 추세이며, 이들을 감지하고 적발하기가 매우 어려운 실정이다. 이로 인해 농산물 유통환경의 공정거래 질서는 침해되고 시장에 대한 신뢰는 훼손되곤 한다. 따라서 거래투명성을 제고하고 유통비리를 구조적으로 개선하기 위한 과학적이고 자동화된 부정탐지시스템의 필요성이 어느 때보다도 절실히 요구되는 상황이다. 본 연구에서는 데이터마이닝의 의사결정나무를 이용하여 실제 발생하지 않은 거래를 실물 없이 거래한 것처럼 조작하여 대금을 정산하는 행위인 허위거래를 탐지하는 모형을 제시하였다. 이를 위해 실제 농산물 도매시장의 데이터를 수집하였고, 데이터의 정제 및 표준화 등의 선행작업을 수행하였다. 또한 변수 간의 상관관계 및 분포도 분석 등을 통해 데이터의 특성을 파악한 후 예측모형을 구축하여 허위거래와 정상거래를 분류하는 패턴을 도출하였으며, 최종적으로 시험용 데이터를 이용하여 모형을 평가하는 단계를 거쳐 결과의 적합성을 확인하였다. 향후 데이터마이닝을 이용한 부정탐지 모형을 허위거래뿐만 아니라 낙찰부정, 경매조작 등과 같이 다양화되는 부정거래에 적용하게 되면 보다 지대한 효과를 거둘 수 있으리라 사료된다.

With the rapid evolution of technology, the size, number, and the type of databases has increased concomitantly, so data mining approaches face many challenging applications from databases. One such application is discovery of fraud patterns from agricultural product wholesale transaction instances. The agricultural product wholesale market in Korea is huge, and vast numbers of transactions have been made every day. The demand for agricultural products continues to grow, and the use of electronic auction systems raises the efficiency of operations of wholesale market. Certainly, the number of unusual transactions is also assumed to be increased in proportion to the trading amount, where an unusual transaction is often the first sign of fraud. However, it is very difficult to identify and detect these transactions and the corresponding fraud occurred in agricultural product wholesale market because the types of fraud are more intelligent than ever before. The fraud can be detected by verifying the overall transaction records manually, but it requires significant amount of human resources, and ultimately is not a practical approach. Frauds also can be revealed by victim's report or complaint. But there are usually no victims in the agricultural product wholesale frauds because they are committed by collusion of an auction company and an intermediary wholesaler. Nevertheless, it is required to monitor transaction records continuously and to make an effort to prevent any fraud, because the fraud not only disturbs the fair trade order of the market but also reduces the credibility of the market rapidly. Applying data mining to such an environment is very useful since it can discover unknown fraud patterns or features from a large volume of transaction data properly. The objective of this research is to empirically investigate the factors necessary to detect fraud transactions in an agricultural product wholesale market by developing a data mining based fraud detection model. One of major frauds is the phantom transaction, which is a colluding transaction by the seller(auction company or forwarder) and buyer(intermediary wholesaler) to commit the fraud transaction. They pretend to fulfill the transaction by recording false data in the online transaction processing system without actually selling products, and the seller receives money from the buyer. This leads to the overstatement of sales performance and illegal money transfers, which reduces the credibility of market. This paper reviews the environment of wholesale market such as types of transactions, roles of participants of the market, and various types and characteristics of frauds, and introduces the whole process of developing the phantom transaction detection model. The process consists of the following 4 modules: (1) Data cleaning and standardization (2) Statistical data analysis such as distribution and correlation analysis, (3) Construction of classification model using decision-tree induction approach, (4) Verification of the model in terms of hit ratio. We collected real data from 6 associations of agricultural producers in metropolitan markets. Final model with a decision-tree induction approach revealed that monthly average trading price of item offered by forwarders is a key variable in detecting the phantom transaction. The verification procedure also confirmed the suitability of the results. However, even though the performance of the results of this research is satisfactory, sensitive issues are still remained for improving classification accuracy and conciseness of rules. One such issue is the robustness of data mining model. Data mining is very much data-oriented, so data mining models tend to be very sensitive to changes of data or situations. Thus, it is evident that this non-robustness of data mining model requires continuous remodeling as data or situation changes. We hope that this paper suggest valuable guideline to organizations and companies that consi

Although the internet is useful for transferring information, Internet auction environments make fraud more attractive to offenders because the chance of detection and punishment are decreased. One of fraud is phantom transaction which is a colluding transaction by the buyer and seller to commit illegal discounting of credit card. They pretend to fulfill the transaction paid by credit card, without actual selling products, and the seller receives cash from credit card corporations. Then seller lends it out buyer with quite high interest rate whose credit score is so bad that he cannot borrow money from anywhere. The purpose of this study is to empirically investigate the factors to detect of the phantom transaction in online auction. Based up on the studies that explored behaviors of buyers and sellers in online auction, bidding numbers, bid increments, sellers' credit, auction length, and starting bids were suggested as independent variables. We developed an Internet-based data collection software agent and collect data on transactions of notebook computers each of which winning bid was over 1,000,000 won. Data analysis with logistic regression model revealed that starting bids, sellers' credit, and auction length were significant in detecting the phantom transaction.