0A Comprehensive Review on Anomaly Detection Techniques for Web Data Logging

The Internet is a dangerous terrain! We hear all the time about websites going down due to denial-of-service attacks or dangerous information on the homepages. Multiple watchwords, dispatch addresses, and credit card data have been blurted into the social sphere in other high-profile cases, vulnerable website druggies to both embarrassment and fiscal threat. Anomaly detection plays a major role in this. It is a step of mining data that detect data points, events, and compliances that diverge from a dataset’s normal geste. Anomalous data indicate critical incidents, similar as specialized glitches, or implicit openings, in case of a change in consumer behavior. Involved dashboards, precisely tuned alert rules and quests through logs, making it a stoner experience. Machine learning has been proposed to improve detection technology for anomalies in recent years, especially in the field of anomalies detection. Here, the authors provide a survey/comparative analysis of published research articles on anomaly detection in correlation with web log analysis. The approach varies, from using traditional machine learning algorithms to two-step algorithms, from neural networks to unsupervised learning and natural language processing to modern methods.