Ai Based Credit Scoring System With Dynamic Risk Assessment

Abstract

Credit cards are now potentially the most popular
mode of payment for both offline and online
purchases thanks to new developments in
electronic commerce systems and
communication technology; as a result, there is
much more fraud involved with such
transactions. Every year, fraudulent credit card
transactions cause businesses and individuals to
lose a lot of money, and con artists are constantly
looking for new tools and techniques to commit
fraud. Researchers face a difficult task when
trying to identify credit card theft since criminals
are quick-thinking and inventive. The dataset
provided for credit card fraud detection is
severely unbalanced, making it difficult for the
system to detect fraud. The use of credit cards is
quite important in today's economy. It is an
essential component of every family, company,
and global enterprise. While using credit cards
responsibly and safely can have many benefits,
engaging in fraudulent behaviour can have a
negative impact on your credit and finances.
There have been several solutions proposed to
address the escalating credit card theft. The
increased use of electronic payments is now
significantly impacted by the detection of
fraudulent transactions. As a result, methods that
are efficient and effective for identifying fraud in
credit card transactions are required. Gradient
Boosting Classifier, a machine learning
methodology, is suggested in this research as a
smart method for identifying fraud in credit card
transactions. The experimental results show that
the suggested approach worked better than other
machine learning algorithms and reached the
maximum accuracy performance, with training
accuracy of 100% and test accuracy of 91%