Comparison of Novel Optimized Random Forest Technique and Gradient Boosting for Credit Card Fraud Detection with Improved Precision
DOI:
https://doi.org/10.47750/pnr.2022.13.S04.0100Keywords:
Anomaly Detection, Credit Card Fraudulent, Internet Security, Gradient Boosting, Machine Learning, Novel Optimized Random Forest Technique.Abstract
Aim: The purpose of this research study is to detect credit card fraud using Novel Optimized Random Forest Technique (NORFT) and Gradient Boosting (GB).
Materials and Methods: The Novel Optimized Random Forest Technique Algorithm uses parallel Decision Tree technique in addition with Random Forest Technique to improve the prediction of Credit Card Fadulants. Total sample size of 40 is used for testing and analysis, based on Gpower statistical analysis tool by considering gpower 0.8. In NORFT used N=20 and in GB used N=20 to measuring the performance of both algorithms.
Result: Novel Optimized Random Forest Technique provides mean precision of 92.52%, and compared with Gradient Boosting algorithm of mean precision is 88.56%. Statistical significance value was fixed as (p>0.05) and obtained 0.477, this shows that NORFT is not statistically significant with alternative hypotheses.
Conclusion: Based on the result, improved precision comparison results show that the efficiency of the Novel Optimized Random Forest Technique is better than Gradient Boosting.
