Classification of Credit Card Transactions Using Machine Learning

In the finance domain the main difficulty faced by the customers is the fraudulency in crediting the amount. Since the evolution of credit cards increased, the frauds on the other hand joined its hand. Previously, many rule based methods were to detect the fraudulent which were not efficient in handling the wide range of variables. But it is necessary to identify the fraud to avoid customer paying unnecessary credit. Because it can be more fascinating and crucial in the security sector, identifying fraud in the payment of credit cards system. To avoid this problem of fraudulent activity in credit card system machine learning model is developed with the three algorithms. Those three algorithms indulged in the model include logistic regression, random forest classifier, bernoulli naive bayes classifier. The efficiency obtained using logistic regression is 96% and with random forest classifiers is about 98% and with naive bayes it is 95%. Thus, the model analyses the best among the three machine learning algorithms.