Web server access log pattern analysis using GWO-based clustering

Abstract

There is a great increase in the amount of information that is available at any given moment, and the internet is a significant source of information. Every day, more people are using the internet than there was the day before. Numerous research is being conducted to reduce the amount of time individuals spend browsing the internet. In the process known as Web usage mining, mining strategies are carried out on a dataset provided by a proxy server to identify the actions of web users. Clustering is an important technique that has many different uses, including the analysis of data from online logs, customer relationship management (CRM), advertising and marketing, scientific diagnostics, and computational biology, amongst many others. Clustering is the collection of data objects that are connected. The problem with clustering is the fact that there are multiple types of measurements that might determine whether or not two things are similar or unique. The paper presents the clustering and classification of web server log access. The clustering is performed using the grey wolf optimization algorithm and classification is performed using five classifiers i.e., support vector machine, random forest, k-NN, decision tree, and gradient boosting. The performance was also compared with existing state of art models and the proposed model has achieved better results.

URN:NBN:sciencein.jist.2023.v11.479