UNCOVERING FRAUDULENT APPS WITH MACHINE LEARNING TECHNIQUES
Abstract
A novel supervised machine learning system is developed to classify network Fraud Application weather it is malicious or benign. To asset the best model in view of detection success rate, a amalgamation of supervised learning algorithm and feature selection method is applied. All the time this study, it is found that the Neural Networks, Random Forest (RANDOM FOREST) based machine learning with wrapper feature selection outperform support vector machine (PCA) technique while classifying network Fraud Application. To evaluate the performance, NSL-KDD dataset is used to classify network Fraud Application using PCA and RANDOM FOREST supervised machine learning techniques. Comparative study shows that the proposed model is more efficient than other existing models with respect to Fraud Application detection success rate. The benign dataset consists of apps that are considered safe and do not pose any threat to users, while the malign dataset contains apps that are considered dangerous and could potentially harm users or their devices. The purpose of this project is to develop machine learning models that can accurately classify apps as either benign or malign based on the features included in the dataset.
Keywords
UNCOVERING FRAUDULENT, MACHINE LEARNING TECHNIQUES