Project title is Distributed Data Mining in Credit Card Fraud Detection. Nowadays, credit card transactions are increasing very quickly and they taken a major share in the US payment system. This will increase the rate of stolen bank account numbers and cause huge loss to the banks. We developed improved fraud detection system to maintain the viability of the US payment system. Before, all banks have used fraud warning systems to protect the US payment system. By using Large-scale data-mining techniques, we can improve the state of the art in commercial practice. Skewed distributions of training data and non-uniform cost per error are the technical problems in fraud-detection task. But these problems are not studied in data mining community and knowledge-discovery. Scalable techniques to examine the huge amounts of transaction data that efficiently compute the fraud detectors in a timely manner is an important problem for e-commerce.   In this paper, we will survey and evaluate the various techniques to solve these three main problems concurrently. The proposed methods of combining the various learned fraud detectors under a cost model are very useful and general. According to our empirical results, by using distributed data mining of fraud models we can significantly reduce the loss due to the fraud. Nowadays with the rapid advances of electronic commerce on the Internet the use of credit cards for purchases has become more famous and convenient. Many people are using credit card to purchase household items. Now Credit card transactions have become an important standard for Web based e-commerce and internet. In the US and other European countries, peoples are making online shopping through credit card. In future, credit card transactions are expected to increase rapidly. But growing number of credit card transactions gives more chance for thieves to steal the credit card numbers and then to make huge loss to banks. When banks lose money due to credit card fraud, credit card holders needs to pay that loss to the bank through higher fees, reduced benefits and higher interest rates. Card holders used old fraud detection methods to reduce the misuse of credit cards. The credit card industry has studied for many years about the computing models for automated detection systems. Recently, these models are the subject of academic research with respect to the e-commerce. The credit card fraud-detection domain gives many challenging issues for data mining. They are given below: • Every day banks will process millions of credit card transactions. To process such massive amounts of data, highly efficient techniques are required. • The data is highly skewed — many more credit card transactions are legitimate than fraudulent. • Typical accuracy – based mining techniques can generate highly accurate credit card fraud. Advertisement: Download projects in studentprojectguide.com Buy this project: