Please read question in comment section,.

Now a days people are interested in ecommerce shopping and many people order the selected products and purchase through ecommerce sites. The payment is online payment. Using users credit card or debit card details and doing the online transactions for purchase of the products. But the main problem is sometimes happening the fraud transactions in ecommerce sites. The cyber attackers apply the multiple types of cyber attacks and stolen the target online users credit card or debit card details and perform the fraud transactions in ecommerce sites. In detection of fraud transactions introduced the methods are using the genetic algorithm. But in this method unable to detect the overall possible fraud transactions and it is time consuming process. So, in proposed system introduced the new method is backlogging method for detection of the fraud online and card transactions in ecommerce websites.

In this proposed project collect the last recent transactions of the user and detect the patterns based on user spent time, location and other information and apply of the location behavior analysis model and trying to detect the fraud user information. In proposed system with backlogging system verify the patterns of the user last three transactions and if any unknown patterns detected in the transactions again perform the reverification for detection of the fraud transactions. If the user detected as fraud user and immediately block the particular user card details and limit and control the fraud transactions in ecommerce websites.

The information or data mining techniques are useful for the detection of the unknown or abnormal patterns in the collected data. The example techniques are neural networking techniques or decision-making techniques or different data mining algorithms are available for detection of the fraud data in the collected input data based on the patterns. The example of the mathematical formulas, machine learning methods are also useful for the detection of the fraud user’s information. Some of the parameters of the information mining are consider in the fraud detection. The parameters are sequence of data, clustering operational details, clustering of the information, forecasting information.

The different techniques of the data mining are used for detection of the credit card fraud transactions are apply the data clustering technique, by apply of the rules are affiliation rules and irregular pattern detection techniques. In some of the methods calculated the threshold value and based on the threshold value like score information categorize the transactions into normal genuine transactions and fraud transactions. The cyber attackers are collecting the target users card information through fake mobile calls, or automated teller machine ATM or through unsecure websites or users web history information like cookies information. So, apply of the information mining techniques not possible to detect the fraud transactions details accurately. It has the some of challenges.

The challenges of data mining are the transactional information is the type of more skewed data. So the detection is not reliable and sometimes due to the duplicate transactions do not get the accurate fraud detection results with data mining methods. For accurate detection of the fraud transactions introduced this proposed system is backlogging technique and pattern behavior analysis.

The main purpose of the proposed system is identification of the fraud users and fraud transactional information from the large input transactional data in ecommerce transactions.

The main objective is detection and prevention of the online fraud transactions in the internet and online web applications.

In this section discuss about the different types of previous methods for detection of the fraud transactions in the static and web applications.

Actually, the fraud transactions are divided into three categories of the credit card or online fraud transactions. If the cyber attackers or hackers illegally gather the customer credit card details from the customer through multiple ways like making of the fraud calls convince the customer and attract the customer with different offers provided to customer and collect the customer card data and making the fraud transactions is called the customer side fraud attractions. The customer is responsible for this type of fraud transactions. For prevention of these types of transactions customer must be maintain the card information securely and do not share the card information to any outside people at any situation.

The second type of frauds are merchant side fraud transactions. Any ecommerce site maintains the database system and payment gateway system. The customers entered payment information is maintained in merchant side database system. If the merchant lost the database information due to the cyber attacks like SQL injection attack or denial of service attacks or any malware security attack, many customers lost the personal and credit card information and hackers may be perform the many fraud transactions through stolen data. In this type of frauds, the merchant is the main responsible person for the customer personal and credit card information and as per security regulations the merchant paid the penalty amount to customers due to violation of the security and privacy rules.

The third type of the fraud transaction is the internet fraud transactions. In this type the hacker or attacker create the malware programs or scripts or content and inject the malware content into user system or user account through spam emails. Some times users get the spam or phishing emails. If any user responds on the phishing email and send any cards information, that is accessed by the attacker illegally and perform the fraud transactions. These are different types of frauds and based on the type of the fraud transaction using the different types of the techniques for detection of the fraud transactions in ecommerce sites.

The credit card information is used for different purposes. In emergency payments using the credit card information. recharge operations, easy payment operations, for getting of loans using the credit cards. In internet and ecommerce sites perform the card not present (CNP) transactions and because of the card not present operations the credit card fraud transactions increased enormously in present society. In this verification of the user identity with strong authentication system is required. But the merchants and ecommerce sites does not use the robust authentication system and due to this lack of access control on the user’s data, happened the more credit card fraud transactions in card not CNP transactions.

In datamining field different types of techniques introduced by different researchers for detection of fraud transactions. The techniques are

The example of this type of technique is the market-based investigation methods. In this method by applying of the association rules identify the relationship between the customer purchased products in the shopping cart of the ecommerce site. By apply of the rules and detect the patterns and trying to detect the duplicate transactions in the same time. Based on the outcomes identify the fraud translations (Jeong, Seong Hoon & Kim, Hana & Shin, Youngsang & Lee, Taejin & Kim, Huy Kang, 2015).

In this technique first collect the input transactional information and prepare the transactional dataset. By apply of the classification and prediction methods classify the common features of the data and irrelevant data. Remove the relevant information from the input transactional data and apply of the prediction technique predict the overall common features of cleaned information and predict the fraud transactions.

In this technique divide the overall hung amount of the transactional information to different types of sub groups of information. The sub groups are called as clusters. In some cluster objects are same and some of the cluster’s objects are different. By using the method is inter class comparison method and compare the transactions on each cluster and identify the fraud transactions with multiple number of clusters and iterations. If clusters are more, the fraud detection accuracy also more.

In this technique identify the possible different patterns in the user input transactions and according to the abnormal patterns identify the fraud transactions in the more input transactions. In pattern analysis consider the time series and time of each transaction and trying to detect the fraud transactions.

First collect the overall credit transactions of the users and apply yhr data analysis models and analyze the data. Using the K-means algorithm classify the data and perform analysis. By apply of K-means cluster the data and categorize the clusters into three categories are low clusters, medium clusters and high clusters. Based on cluster category identify if the clusters having more suspicious transactions trying to identify the users of those transactions and verify the users information and detect the fraud transactions and users details also.

In genetic algorithm apply the clusters and after clustering calculated the threshold score value. The score value is used for classification of the transactions into relevant and irrelevant type. Provide the conditions with the minimum threshold value. If the condition satisfied the translation is correct and if the condition not satisfied that transactions are considered as the fraud transactions. Based on the credit card usage time, credit card limit and balance identify the fraud transactions with iterated clustering process. Every time created the new clusters with previous clusters of data and trying to detect the clusters (Modi, Krishna, 2017).

In proposed system first stored the customers transactions information and maintain the logger system with backlogging technique. In this method maintain the previous transactions of the customers and current transactions also. If user enter the credit card information more than 3 times, the proposed ecommerce website shows the alert message about the blockage of the particular user credit card and internally start the verification of that customer transactions automatically by the admin user.

In verification process the admin first verify the identity of the user with the authentication method. Check the login details of the user and once the authentication is done, next verify the recent transactions of the user. Next consider the user characteristics like user performed transaction time, credit card number, location and other characteristics. This is user friendly application and the ecommerce website usage of thee customers easy to operate and easy to perform the secure credit card fraud transactions with this proposed application.

The system architectural diagram is

Figure 1: System Architecture diagram

In above architectural diagram the application users customer or user and administrators are interact with the ecommerce website and perform the operations. The customers data is maintained in database system and ecommerce site stored the information in database system.

The technical request and response process diagram is

Figure 2: Internal Working Process diagram

The above diagram shows the technical information. The ecommerce website is running in web browser. The user or admin users send the request to web server through controller. The controller will process the request and the controller interact with the database system and get the required information from the database and send the proper response to user in web browser.

Any application system design is developed with the UML diagrams. The UML diagrams are different types. Some of the UML diagrams of the proposed system is

In use case diagram create the user actions or use cases and the overall system user’s functionalities are described. The below diagram is proposed system use case diagram. The users of the proposed system is admin and customer users. The admin use cases and customer use cases are provided in round ovel symbols are called use cases. The links between the users are called actors and use cases is relation links. The use cases are login, registration, forget password, view customers, add card details, view orders and view fraud orders, search product, order product, payment and it include cash and card payments.

Figure 3: Use case diagram

The proposed system having the actors are admin and customer. Both roles of the users having the sequential operations. The operations are described in the below sequential diagram. The life lines of the sequential diagram is login, view customers and transactions, card data, fraud and order details.

Figure 4: Sequence diagram

The communication diagram is created based on sequential diagram. The proposed system users communication described with the communication diagram in detailed type.

Figure 5: Communication diagram

The activity diagram is described the overall activities of the proposed system users and it show the start and stop of the application activities. The rectangle symbol is activities, Rambus symbol for condition checking and it include the start and stop symbols also.

Figure 6: Activity diagram

The proposed system is implemented by two types of roles of users are customer, admin users. The ecommerce website admin is admin user and website used by the web users are customers. The customer module, user module, fraud detection module. These three modules implemented in the proposed system.

In implementation of the system first the admin login ecommerce website. Developed the login and registration screens in website. By enter of the admin user details like login email id and password login into system. verified the admin login details with database information and if the user is authenticated successfully, move to the admin dashboard. The home page of the admin user having the activities like view the customers data, add the customer credit card or debit card data in website, after customer request for order of the product. Admin can view the customer recent orders and transactions. If any customer is fraud, the frauds information can check by admin.

The customer is the website user and register into web site. After registration enter login data and check the login authentication of customer with database data. Once authentication is correct, move to the customer dash board or home page. The customer can search the products with search facility and if any products need, order the products and paid the amount in payment mode. In payment select the type of payment, if card type select credit or debit card payment and if the customer enter wrong card details more than three times, the customer card is blocked by website admin and the admin background check the customer is fraud or original customer with recent 3 transactional history is backlogging history and if valid unlock the card, if the customer is fraud blocked the card and send the notification to corresponding customer about the fraud transactions.

The output results of the proposed system screens are

Login Page

Figure 7: Login

Register Page

Figure 8: Registration

Admin Home

Figure 9: Admin Home

Customer Home

Figure 10: Customer Home

Fraud Detection View Page

Figure 11: View fraud details

In ecommerce website using this proposed system technique provide the privacy and security for the customer card payment transactions and successfully prevent the fraud transactions in the ecommerce web application with backlogging implementation and blocked the fraud customers card information also.

In future develop the extension for this application is multi factor authentication. By using additional customer verification with multi factor authentication system or random password creation with cryptography techniques and overcome the unauthorized users access problems in the ecommerce web application.

References Jeong, Seong Hoon & Kim, Hana & Shin, Youngsang & Lee, Taejin & Kim, Huy Kang. (2015). A Survey of Fraud Detection Research based on Transaction Analysis and Data Mining Technique. Journal of the Korea Institute of Information Security and Cryptology. . 25. 1525-1540. 10.13089/JKIISC.2015.25.6.1525. . Modi, Krishna. (2017). Review on fraud detection methods in credit card transactions. . 10.1109/I2C2.2017.8321781. .