Description: Field of Invention
With the help of data mining, it is possible to detect and prevent financial statement fraud. Data mining is a technique that seeks to identify hidden patterns and correlations in large datasets. This data mining ability was put to use in order to protect against the possibility of fraudulent financial reporting. Big data approaches include statistical, machine learning, data mining, and optimization techniques, all of which are classified as such. Data envelopment analysis (DEA) is a technique that is well-known for its application in the context of big data analytics. Diffractional linear programming (DEA) is a technique that is widely used to discover the best practises for firms given a set of performance metrics. Supply chain and environmental research have both embraced the use of DEA for big data sets for performance evaluation, particularly in supply chain research. However, it was suggested that DEA be utilised as a tool for auditing in order to identify potential audit targets, which was approved by the committee. In prior study, DEA has also been used to extract value from large amounts of data. It is possible to investigate the amount and velocity of big data by applying DEA as an auditing tool to detect anomalies in a large data set of data.
Background of the Invention
As a result of the rising use of credit card transaction services through a number of channels, massive volumes of data are being generated and stored in data warehouses on a daily basis. The knowledge and insight gathered from this database can provide an operator with a competitive advantage in the fight against fraud, according to the researchers. As a result, payment card fraud has risen to the top of the priority list for most financial institutions. Financial institutions' customers' confidence in the transaction security provided by the service provider has been eroded as a result of fraud, (US9165051B2)which represents a significant source of lost revenue for them. As a result of increased competition and impairments, fraud has gone from being a source of concern for delivery services to being a non-issue. Effective analysis and fraud detection systems for card-based financial institutions can help prevent a major loss, save dollars, and restore client confidence in the security of their transactions, according to the Financial Services Institute. When fraud is detected automatically, it is feasible to take action against the perpetrators of the scam, such as service refusals and measures against the fraudsters. As seen by the vast amount of online transactions, identifying and analysing fraud is a time-consuming and complex task(US9836455B2). As a general rule, the more complex and diversified the delivery channels for financial services are, the more prone they are to fraudulent activity. Institutions will have to react fast in the future if they wish to stay up with fraudulent users and their new problems. However, even if rules-based systems with precise criteria for certain traits may be capable of dealing with some types of fraud, they are unable to keep up with the endless number of new possibilities that are always being presented. Alternatively, fraudsters might simply alter their strategies in order to avoid detection.
A machine learning selection programme develops and picks the rules that are normalised, altered, and stored by the software programmes, which are then used by the software programmes to make decisions. Based on the discovered fraud rules, this module creates a set of profilers as well as a set of profiling templates that are instantiated in response to the fraud rules' conditions(US8271403B2). The storage component will store information for a number of functions, such as user profiles, training and testing data for the detector module, and so on. Detector modules are provided with training data, and they are trained to recognise positive and negative patterns in response to the training data. The profiling module will be used to construct user profiles based on the information provided by your fraud rules and the template set that was provided. A profile for each user will include information on the cardholder's buying habits and other transactional information such as the frequency of transactions, the average amount spent per transaction, the location of purchases, and other factors. These profiles can be used to compare new transactions with user profiles, assess if there is a significant deviation from the user profile, and offer numeric values for each fraudulent signal that is detected(US9760656B2). This evidence is then combined by the detector, which produces an output proposal for the new transaction based on the numeric values contained in the evidence components. Our technology is simple to use and can be implemented by any organisation that processes real-time online credit card transactions. Artificial Neural Network (ANN): The proposed architecture comprises a training model known as the Artificial Neural Network (ANN), which will be capable of accurately classifying whether an incoming transaction is fraudulent or not in real time. A warning will be displayed on the Fraud Manager's dashboard or their authorised account when the anticipated transaction is found to have a Fraud Score of 0.7 in real time. This is made possible thanks to the model's interaction with Apache Airflow, which allows the prediction model to be automated.
Summary of the Invention
A major purpose of our investigation is to learn more about the procedures that banks and other card issuers have in place to combat the growing and alarming challenges that are linked with payment card transactions. We'll also talk about the security features of a payment card, as well as how to keep payment cards safe from fraudsters when doing both online and offline transactions. It doesn't matter what type of fraudulent conduct is being investigated because fraud detection and prevention systems are capable of quickly adapting to changing circumstances and identifying new fraud trends in real time. Create a data mining framework to detect financial statement fraud. Create a set of financial statement fraud prevention and detection rules. Financial and economic fraud is a significant issue for banks and other financial institutions. Fraud can lead to a decline in industry trust, an unstable economy, and an increase in the cost of everyday necessities. Traditional fraud detection methods relied on manual techniques such as reviewing, which are ineffectual and problematic because of the challenges that arise in this issue. This situation is thorny. Because of their ability to spot little irregularities in large informational indexes, computational intelligence (CI) approaches and techniques and data mining methods have proven to be useful. Various data mining strategies are explored in the research for detecting fraud, however there isn't a particularly effective approach in the credit card literature that outperforms them all.
Brief Description of Drawings
Figure 1: Work flow of the Proposed Fraud detection method
Figure 2: Proposed Fraud detection Process.
Detailed Description of the Invention
Theft of the physical card or negotiation of the card information and/or cardholder information is the beginning of a credit card fraud. As simply as a clerk at a retail establishment copying sales receipts, the compromise can occur. Security failures on databases storing credit card information can be particularly extensive and costly due to the geographic reach that they may entail due to the increasing rise of credit card use worldwide and particularly through the Internet. At least 40 million credit card accounts were stolen from one database in 2005 after a single breach of an extremely large database containing credit card data. New cards had to be issued to millions of card holders around the world as a result of this.
Even though cardholders are fast to report lost or stolen cards, a compromised account might go weeks or months without any fraudulent activity being detected, making it difficult to pinpoint the source of the compromise. If the cardholder receives a request statement just once a month, he or she may not notice fraudulent use until that statement arrives.
Tools such as data mining and data analysis are employed in the creation of new security measures and authentication procedures in order to detect evolving fraud developments in transactional databases maintained by financial institutions. In response to enhanced security measures, fraudsters modify their strategies while continuing to use the current service platform to carry out unlawful activities. It is common for fraud trends to be identified after new security measures are implemented, resulting in the need for an organization's fraud prevention strategy to be reorganised. Undoubtedly, one of the most important objectives of these detection systems is to identify patterns of transactions and applications that may be fraudulent or suspicious in nature. Transactional fraud occurs when fraudsters make extra purchases using a genuine credit or debit/prepaid account that has previously been established by the victim. To counteract those who attempt to perpetrate fraud of any kind by exploiting security flaws, researchers are on a never-ending quest for effective countermeasures to use. When it comes to this environment, fraud protection and detection technologies have arisen as a vital requirement. These tactics are effective because they are flexible enough to adjust to changing fraudster behaviour. We took a step forward in this area because inline financial fraud prevention and detection continues to be a difficult problem for researchers to solve.
As a result, a thorough evaluation of present ways to detecting financial and economic fraud is now required. Data mining can uncover previously unknown patterns and predict future market trends and practises. In order to gain an advantage, organisations can make proactive and information-driven decisions. Numerous financial applications have been linked to data mining, including the development of exchange model predictions and the appraisal of loans, as well as portfolio enhancement and fraud detection and bankruptcy prediction. In today's world, fraud detection is one of the most critical areas of data mining. Detection of fraud by screening is made more efficient by automating the laborious process of identifying frauds. data mining and data analysis are employed in the creation of new security measures and authentication procedures in order to detect evolving fraud developments in transactional databases maintained by financial institutions. In response to enhanced security measures, fraudsters modify their strategies while continuing to use the current service platform to carry out unlawful activities as architecture shown in Fig:1. Fraud should be prevented and detected by utilising data mining and artificial intelligence capabilities. Our efforts in the identification and prevention of card-based application fraud go hand in hand with one another shown in Fig:2.
4 Claims & 2 Figures , Claims: The scope of the invention is defined by the following claims:

Claim:
1. The Novel Fraud detection Technique using Data Mining comprising the steps of:
a) Examines and surveys the present use of data mining techniques for financial statement fraud detection.
b) Investigated the amount and velocity of big data by applying DEA as an auditing tool to detect anomalies in a large data set of data and safeguard the fraud detection system at the account and transaction levels.
c) Presented a simplified fraud detection model that includes the benefits of existing safe transaction processes and Payment card holders will benefit from more reliable and secure services.
2. The Novel Fraud detection Technique using Data Mining as claimed in claim 1, fraudulent conduct is being investigated because fraud detection and prevention systems are capable of quickly adapting to changing circumstances and identifying new fraud trends and create a data mining framework to detect financial statement fraud. Create a set of financial statement fraud prevention and detection rules.
3. The Novel Fraud detection Technique using Data Mining as claimed in claim 1, investigated the amount and velocity of big data by applying DEA as an auditing tool to detect anomalies in a large data set of data and safeguard the fraud detection system at the account and transaction levels.
4. The Novel Fraud detection Technique using Data Mining as claimed in claim 1, presented a simplified fraud detection model that includes the benefits of existing safe transaction processes and Payment card holders will benefit from more reliable and secure services.