Creating Unbiased Machine Learning Models by Design

Unintended bias against protected groups has become a key obstacle to the widespread adoption of machine learning methods. This work presents a modeling procedure that carefully builds models around protected class information in order to make sure that the final machine learning model is independent of protected class status, even in a nonlinear sense. This procedure works for any machine learning method. The procedure was tested on subprime credit card data combined with demographic data by zip code from the US Census. The census data serves as an imperfect proxy for borrower demographics but serves to illustrate the procedure.

Behavioral Based Credit Card Fraud Detection

Credit card fraud is a significant threat in the BFSI sector. This credit card fraud detection system analyzes user behavioral patterns and their location to identify any unusual patterns. This consists of user characteristics, which includes user spending styles as well as standard user geographic places to verify his identity. One of the user behavior patterns includes spending habits, usage patterns, etc. This system deals with user credit card data for various characteristics, which includes user country, usual spending procedures. Based upon previous transactions information of that person, the system recognizes unusual patterns in the payment method. The fraud detection system contains the past transaction data of each user. Based on this data, it identifies the standard user behavior patterns for individual users, and any deviation from those normal user patterns becomes a trigger for the detection system. If it detects any unusual patterns, then user will be required to undergo the security verification, which identifies the original user using QR code recognition system. In case of any unusual activity, the system not only raises alerts but it will block the user after three invalid attempts.

RECOGNITION OF IDENTIFICATION DATA OF BANK CARDS

This article describes the recognition of bank card information. Recognizing an object with a camera is one of the most important tasks at the moment. Recognizing credit card data at the same time is a rather complex algorithmic task, but at the moment the implementation of this task is very relevant and in-demand due to the increase in the number of payment transactions via mobile devices. The implementation of this task can save a person from having to enter most of the data when making online payments. The fundamental difficulties of this problem are discussed and methods for solving it are proposed. The problem under consideration is solved for the case of application on mobile devices, which imposes strict requirements for computational complexity. The article presents the results of a formal analysis of the performance and accuracy of the proposed algorithm. The error spectrum of the recognition system as a whole shows that the proposed algorithm solves the problem with the required accuracy. The main question that was investigated at this work: is it possible to use the Tesseract OCR library for text recognition from video images, for example, timecode? That is, digital time data embedded in the footage images. This is important for the automation of individual procedures for video technical expert studies. Object recognition by the camera is one of the most important tasks at the moment. The fundamental difficulties of this problem are discussed and methods for its solution are proposed. The article presents the results of a formal analysis of the performance and accuracy of the proposed algorithm. The spectrum of errors of the recognition system as a whole shows that the proposed algorithm solves the problem with the required accuracy.

The COVID-19 Shock and Consumer Credit: Evidence from Credit Card Data

We use credit card data from the Federal Reserve Board's FR Y-14M reports to study the impact of the COVID-19 shock on the use and availability of consumer credit across borrower types from March through August 2020. We document an initial sharp decrease in credit card transactions and outstanding balances in March and April. While spending starts to recover by May, especially for risky borrowers, balances remain depressed overall. We find a strong negative impact of local pandemic severity on credit use, which becomes smaller over time, consistent with pandemic fatigue. Restrictive public health interventions also negatively affect credit use, but the pandemic itself is the main driver. We further document a large reduction in credit card originations, especially to risky borrowers. Consistent with a tightening of credit supply and a flight-to-safety response of banks, we find an increase in interest rates of newly issued credit cards to less creditworthy borrowers.

Changes in Consumer Behaviour in the Post-COVID-19 Era in Seoul, South Korea

To prevent the spread of COVID-19, the Korean government promoted strong social distancing policies and restricted the use of confined areas and spaces that are likely to cause widespread infection, including religious facilities. The policies affect the consumption behaviours of Korean citizens. The purpose of this study is to examine changes in the consumer behaviours of citizens following the outbreak of COVID-19 in South Korea. Using credit card data from January to June 2020 in Seoul, this study examines the changes in consumption by industry type. Consumption types were classified into education, wholesale and retail, online purchases, food service, leisure, and travel. The industry that was most affected was the travel industry, which did not recover following the decline in consumption due to COVID-19. To examine consumer changes in credit card transactions due to widespread infection, a correlation analysis was conducted between the amount of consumption according to credit card transaction data (card consumption) and the number of confirmed patients and policy implementation by step. For more detailed analyses, group infections in the Guro-gu and Yongsan-gu neighbourhoods were investigated. In Guro-gu, no significant results were found in the area experiencing massive group infection. In Yongsan-gu, a significant negative correlation in consumption and number of cases was found in Itaewon 1-dong, an area with mass infection, and a positive correlation was found in the surrounding areas. Nevertheless, no significant correlations between changes in consumer behaviours and effects of COVID-19 were found as a result of the analysis herein.

Credit Card Fraud Detection and Prevention using Machine Learning

This research focused mainly on detecting credit card fraud in real world. We must collect the credit card data sets initially for qualified data set. Then provide queries on the user's credit card to test the data set. After random forest algorithm classification method using the already evaluated data set and providing current data set[1]. Finally, the accuracy of the results data is optimised. Then the processing of a number of attributes will be implemented, so that affecting fraud detection can be found in viewing the representation of the graphical model. The techniques efficiency is measured based on accuracy, flexibility, and specificity, precision. The results obtained with the use of the Random Forest Algorithm have proved much more effective