Online Credit Card Application and Identity Crime Detection

Online Credit Card Application and Identity Crime Detection

Ramkumar.E & Mrs Kavitha.P

School of Computing Science, Hindustan University, Chennai

ABSTRACT

The credit cards have found widespread usage due to the convenience they offer. Credit applications are Internet or paper-based forms with written requests by potential customers for credit cards, mortgage loans, and personal loans. The credit application fraud is a specific case of identity crime. The credit application fraud pattern is represented by a sudden and sharp spike in duplicates within a short time, relative to the established baseline level. The existing communal detection and spike detection . CD finds real social relationships to reduce the suspicion score, and tamper resistant to synthetic social relationships. It is white list-oriented approach on a fixed set of attributes. SD finds spikes in duplicates to increase the suspicion score, and is probe-resistant for attributes. The existing system detects the whether the applicant is fraud. It is the attribute-oriented approach on a variable-size set of attributes. In the existing system the fraudster datum are stored in the database manually. In this proposed system, CD and SD can detect more types of attacks, better account for changing legal behavior, and remove the redundant attributes and to store the fraudulent datum in blacklist using CBR algorithm. CBR algorithm analysis using retrieval, diagnosis and resolution to make the data more secure and to find the fraudulent data. The data that already present or fraudulent is encountered and thrown into the blacklist. Together CD, SD and CBR ensure the data provided by the customer is original. This proposed system makes the system more efficient and enhance the security.

Keywords: Communal Detection, Spike Detection, Case Based Reasoning, fraud detection

1. Introduction

   The data mining consists of multiple detection algorithms. Data mining detection algorithm are used in the online credit card application. The algorithms are used in this system is the spike detection, communal detection and CBR algorithms. These algorithms are used to detect the fraud and throw the data in the database as original data or blacklist database. This system updates the database manually. This system does not give a chance to fraudsters in credit card application. Identity crime is defined as broadly as possible in this system. At one extreme, real identity theft refers to illegal use of innocent peoples complete identity details. These can be harder to obtain

   (although large volumes of some identity data are widely available) but easier to successfully apply. In reality, identity crime can be committed with a mix of both synthetic and real identity details. Credit applications are Internet or paper-based forms with written requests by potential customers for credit cards, mortgage loans, and personal loans. Credit application fraud is a specific case of identity crime, involving synthetic identity fraud and real identity theft. As in identity crime, credit application fraud has reached a critical mass of fraudsters who are highly experienced, organized, and sophisticated. There are two types of duplicates: exact (or identical) duplicates have the all same values; near (or approximate) duplicates have some same values (or characters), some similar values with slightly altered spellings, or both. In short, the new methods are based on White-listing and Detecting spikes of similar applications. White-listing uses real social relationships on a fixed set of attributes. This reduces false positives by lowering some suspicion scores. Detecting spikes in duplicates, on a variable set of attributes. This increases true positives by adjusting suspicion scores appropriately.

   1. Data Mining

      Data mining, the extraction of hidden predictive information from large databases, is a powerful new technology with great potential to help companies focus on the most important information in the data warehouses. Data mining tools predict future trends and behaviors, allowing businesses to make proactive, knowledge-driven decisions. The administrator verifies the provided data with the existing datum to find whether it is fraudulent or original. If the data is original, it will be added to the database otherwise it will be thrown into the blacklist.

   2. CBR Algorithm

      Case-based reasoning (CBR) is now making a significant contribution to the task of fraud detection. CBR systems are able to learn from sample patterns of credit card use to classify new cases, and this approach also has the promise of being able to adapt new patterns of fraud as they emerge. The CBR

      system is the application of adaptive and hybrid learning systems. The CBR problems are previously considered too dynamic, chaotic, or complex to accurately model.

   3. Problem Statement

      The online credit card application is internet based forms. This system detects the fraud applicant using the data mining algorithms. The existing system used two algorithms they are spike detection and communal detection algorithm. These two combine together detects whether the applicant is fraud or original. The proposed system combing with these algorithms CBR uses to find the fraudulent data and puts it into the blacklist. This system uses for the match analysis from the existing blacklist database to make the system efficient and secure

   4. Objective of the work

Data mining is concerned with analysis of large volumes of data to automatically discover interesting regularities or relationships which in turn leads to better understanding of the underlying processes. The data mining consists of multiple algorithms, some algorithms uses for the detection of fraud detection in credit card. Online credit card application uses these algorithms communal and spike detection uses to detect the multiple applicant and with the artificial intelligent CBR algorithm uses to make the fraudulent data in the black list.

2. Existing System

The credit card application the system detects whether the applicant is fraud or original. The existing system implements two data mining layers they are communal and spike detection. The Communal Detection is used to find the suspicious data of the fraudulent people. It also used to find the communal relationship that are near to reflect the family bond.(i.e parent child). It is white list oriented. The spike detection, it is attribute oriented. This does not detect the fraud but updates the system regularly and attributes regularly other than than the communal detection.

2.1 Demerits of existing system

The system detects the whether the data is fraud or original. If the system is data is fraud the process do not proceed to the next level.

The system is attribute oriented that the data is updated in the communal detection manually.

The system does not verify from the blacklist database. Through the spike detection the system updates the attributes regularly.

The system is not secure and it detects the original data also as fraud. (for eg.- twins applying the card is also detects as the fraudulent data).

1. Proposed Method

   The communal detection focused on attacks in the white list by fraudsters when they submit applications with synthetic relationship. The volume and ranks of the white lists real communal relationships change over time, to make the white list exercise caution with (more adaptive) changing legal behavior, the white list is continually being reconstructed. The spike detectin is attributes oriented. It cannot be detected by fraud attribute will be updated regularly. The attributes used in spike detection will not be communal detection. By using the spike detection and communal detection detects the fraudsters in credit card application In addition to communal detection and spike detection we use case based reasoning algorithm to make this approach more efficient. CBR implements retrieval, diagnosis and resolution to make the data more secure. The CBR used to analyze and retrieval of data from the existing blacklist. The fraudulent datum is moved to the blacklist and the original datum is stored in the database.

   1. Merits of Proposed Method

      With the existing algorithms proposed the Case based reasoning algorithm to make the system secure. This system verifies the data with blacklist data. The blacklist data is verified with the CBR algorithm that is to used to find the fraudulent data.

      This system updates the database automatically by using the data mining algorithm.

2. Fraud Detection Algorithms

   1. Communal Detection

      If there are two credit card applications that provided the same postal address, home phone number, and date of birth, but one stated the applicants name to be John Smith, and the other stated the applicants name to be Joan Smith. Either it is a fraudster attempting to obtain multiple credit cards using near duplicated data. Possibly there are twins living in the same house who both are applying for a credit card. Or it can be the same person applying twice, and there is a typographical error of one character in the first name. It is crucial because it reduces the scores of these legal behaviors and false positives. There are two problems with the white list. First there can be focused attacks on the white list by fraudsters when they submit applications with synthetic communal relationships Second, the volume and ranks of the

      white lists real communal relationships change over time. To make the White list exercise caution with (or more adaptive to) changing legal behavior, the white list is continually being reconstructed.

   2. Spike Detection

      SD finds spikes to increase the suspicion score, and is probe resistant for attributes. Probe resistance reduces the chances a fraudster will discover attributes used in the SD score calculation[3]. It is the attribute- oriented approach on a variable-size set of attributes. The redundant attributes are continually filtered; only selected attributes in the form of not-too-sparse and not-too-dense attributes are used for the SD suspicion score.

   3. CBR Algorithm

      1. Retrieval

         Nearest neighbor matching is common to many CBR systems. Again using the basic exploratory facilities of CBR test bed, a set of cases which were considered to be very similar, i.e[12]., above a certain percentage of similarity, were retrieved.

      2. Diagnosis

         Applying the general principle of threshold retrieval, a multi-algorithmic approach to final match analysis was developed as a result of the design and testing of a variety of single discrimination algorithms.[12] It has been suggested that no single algorithm may perform equally well on all search and classification tasks, and that an algorithms improved performance in one learning situation may come at the expense of accuracy in another.

      3. Resolution

         If more than one algorithm is asked to diagnose the set of cases retrieved for an unknown credit request, it is possible that the algorithms may disagree on the result, and resolution strategies were implemented to resolve the varying diagnoses into a single result.

3. System Architecture

   The architecture diagram represents the overall structure of the system. The data is detected for the crime detection using the data mining algorithm communal detection and spike detection algorithm. These two algorithms combine together to remove the negative false and then proceeded to the proposed system algorithm (i.e) CBR algorithm. This algorithm retrieved and diagnosis the datum. If the data is fraud it is thrown into the black list database.

   If the data is original the data is stored in the database.

   The communal detection focused on attacks in the white list by fraudsters when they submit applications with synthetic relationship. The volume and ranks of the white lists real communal relationships change over time, to make the white list exercise caution with (more adaptive) changing legal behavior, the white list is continually being reconstructed. The spike detection is attributes oriented. It cannot be detected by fraud attribute will be updated regularly. The attributes used in spike detection will not be communal detection. By using the spike detection and communal detection detects the fraudsters in credit card application In addition to communal detection and spike detection we use case based reasoning algorithm to make this approach more efficient. CBR implements retrieval, diagnosis and resolution to make the data more secure. The CBR used to analyze and retrieval of data from the existing blacklist. The fraudulent datum is moved to the blacklist and the original datum is stored in the database.

   1. Crime Detection

      The crime detection consists of the two algorithms, communal detection and spike detection. The communal detection detects the fraudsters. This detection is the relationship oriented. This detection is attribute oriented. The spike detection detects the system fraudsters by updating the system attributes. These system finds the data whether the data is original or not. These two detections are mainly involved in existing crime detection.

   2. Finding Legtimate User

      The CBR is used is the fraud detection system that the data is original or not that the data is original or not by retrieving the data from the blacklist verification. This system finds the fraudulent data by the artificial intelligence. The CBR algorithm involves with the data mining concept with match analysis

   3. Blacklist Verification

      With the provided sets of details are taken into consideration to avoid the identity crime. The data is verified using the above algorithms to make the credit card application enormously efficient. If the data is original further processes will be enforced or otherwise the data will be found as fraud and it will be enrolled in the black list.

      Communal detection

      Communal detection

      Crime detection

      Crime detection

      Applicant data

      Crime detection

      Crime detection

      Case base reasoning

      Case base reasoning

      Retrieval, Diagnosis, resolution

      Retrieval, Diagnosis, resolution

      Original data

      Original data

      5.1 System Architecture Diagram

4. Conclusion

   The system detects the fraud detection online credit card application. This system is used avoid the duplicates from the fraudsters while applying the credit card. Data mining algorithms are used this system. The existing algorithm communal detection and spike detection used to detect the multiple applicants. In proposed system combing with the existing algorithm spike detection and communal detection the CBR algorithm is used to make the system more efficient and secure. The CBR algorithm is used to throw the fraudulent data in the blacklist and retrieve the datum from the blacklist database. The identity thief has limited time because innocent people can discover the fraud early and take action, and will quickly use the same real identities at different places.

5. Future Enhancements

The detection of credit card application is used with the dataming layers. This system is used only on the application, in future the fraud detection in credit process (i.e) the card is used by the unauthorized user. This system can be developed with the dataminig system and the with the help of the biometri system.

REFFRENCES

1. A. Bifet and R. Kirkby Massive Online Analysis, Technical Manual, Univ. of Waikato, 2009.

2. R. Bolton and D. Hand, Unsupervised Profiling Methods for Fraud Detection, Statistical Science, vol. 17, no. 3, pp. 235-255,2001.

3. P. Brockett, R. Derrig, L. Golden, A. Levine, and

   M. Alpert, Fraud Classification Using Principal Component Analysis of RIDITs,The J. Risk and Insurance, vol. 69, no. 3, pp. 341-371, 2002, doi: 10.1111/1539-6975.00027.

4. R. Caruana and A. Niculescu-Mizil, Data Mining in Metric Space: An Empirical Analysis of Supervised Learning Performance Criteria, Proc. 10th ACM SIGKDD Intl Conf. Knowledge Discoveryand Data Mining (KDD 04), 2004, doi: 10.1145/1014052.1014063.

5. P. Christen and K. Goiser, Quality and Complexity Measures for Data Linkage and Deduplication, Quality Measures in DataMining, F. Guillet and H. Hamilton, eds., vol. 43, Springer, 2007, doi: 10.1007/978-3-540-44918-8.

6. C. Cortes, D. Pregibon, and C. Volinsky, Computational Methods for Dynamic Graphs, Computational and Graphical Statistics,vol. 12, no. 4, pp. 950-970, 2003, doi: 10.1198/1061860032742.

7. Experian. Experian Detect: Application Fraud Prevention System, Whitepaper, http://www.experian.com/products/pdf/experian_dete ct.pdf, 2008.

8. T. Fawcett, An Introduction to ROC Analysis, Pattern Recognition Letters, vol. 27, pp. 861-874, 2006, doi: 10.1016/j.patrec. 2005.10.010.

9. A. Goldenberg, G. Shmueli, R. Caruana, and S. Fienberg, Early Statistical Detection of Anthrax Outbreaks by Tracking Over-the-Counter Medication Sales, Proc. Natl Academy of Sciences USA (PNAS 02), vol. 99, no. 8, pp. 5237-5240, 2002.

10. G. Gordon, D. Rebovich, K. Choo, and J. Gordon, Identity Fraud Trends and Patterns: Building a Data-Based Foundation forProactive Enforcement, Center for Identity Management and Information Protection, Utica College, 2007.

11. Multiple Algorithms for Fraud Detection,Richard Wheeler and Stuart Aitken, Artificial Intelligence Applications Institute,The University of Edinburgh,

    80 South Bridge,Edinburgh EH1 1HN, Scotland, 2006.

12. Schaffer. C, A Conservation Law for Generalized Performance,2009.

13. O. Kursun, A. Koufakou, B. Chen, M. Georgiopoulos, K. Reynolds, and R.Eaglin, A Dictionary-Based Approach to Fast and Accurate Name Matching in Large Law Enforcement Databases, Proc. IEEE Intl Conf. Intelligence and Security Informatics (ISI 06), pp. 72-82, 2006, doi: 10.1007/11760146.