An Innovative Approach to Detecting Fraud in E-Commerce Applications based on Anomaly Intrusion Detection Systems Using ALO-ELM Approach

Online shopping and selling has been increasingly popular in recent years. Along with the expansion of online transactions, the number of attacks targeting the safety of computer networks has also grown. In recent years, e-commerce sites and online auctions have seen an uptick in fraudulent transactions. Some of these bogus online purchases are the product of hackers breaking onto these platforms. The problem of application-based attacks in e-commerce remains unsolved, even though these facts have received a lot of attention. Model training, feature selection, and data preprocessing must all be executed in this particular order. Prior research on data sample analysis indicates that there are two possible issues with preprocessing that require fixing: multicollinearity and discretization. The study on feature selection aims to reduce noise, improve feature matrix differentiation, and eliminate mutual interference. Feature retrieval is the first step in training ALO-ELM models. When put side by side with two cutting-edge approaches, ALO and ELM, the suggested strategy comes out on top. Accuracy improved by 95.36 percent following application of the approach.