Pecuniary Fraud Uncovering with Intelligent and Collaborative Practices

Abstract

The increasing likelihood of financial fraud has become a major worry in a world where mobile connections are necessary for transmitting substantial volumes of data. The Resnet Autoencoder Lasso Regression Feature Selection Hybrid Model (RALFSHM) is a novel artificial intelligence technique designed specifically for processing financial transaction data. Our artificial intelligence method adopts a methodical approach to tackle the growing risk of financial fraud, which poses a significant threat to both financial institutions and their clients. We start the procedure with data preprocessing using a standard scalar, resolving disparity in the data using the Synthetic Minority Over-Sampling Technique Edited Nearest Neighbors (SMOTEENN). Extracting features applies an artificial intelligence ensemble technique that blends autoencoders, Resnet, and Lasso Regression to acknowledge crucial data patterns. At the same time, the Gradient Boosting Machine and Extreme Gradient Boosting hybrid voting classifier (GXVM) evaluate the model’s performance. The GXVM model hyperparameters are a key component of our artificial intelligence classification task. We fully interpret our model using a dataset of pecuniary transactions. This revolutionary study on artificial intelligence promises improved surveillance and competence in fiscal transactions, marking a breakthrough in the continuous conflict against financial delinquency.