Web User Access Log Analytics Using Neural Learning, Regression and Logit Boost Clustering Techniques for Accurate User Behavioural Pattern Identification
- Title
- Web User Access Log Analytics Using Neural Learning, Regression and Logit Boost Clustering Techniques for Accurate User Behavioural Pattern Identification
- Creator
- R, Gokulapriya
- Contributor
- R, Ganesh Kumar
- Description
- Web Usage Mining (WUM), is the process of mining user behaviour patterns from huge log fles. Weblogs provide substantial input to learning the identity of an online user. Analysis of these patterns extracted from the weblog datasets is currently being explored by various researchers. Due to the recent advent of automation, mining patterns from weblogs are automated. These automated mining processes focus on browsing habits and usage patterns. To make this process of gathering better, there are many ways to look at how users act and put them into relevant groups.Identifying, detecting, and classifying features that demarcate specifc traits that are related is an important task. Conventional research is designed to discover web usage mining strategies through clustering and classifcation methods. However, there is a need to focus on and improve the accuracy of the prediction systems that classify acquired features to fgure out the patterns of web users. Deep learning methods are used to mine weblog data to improve accuracy and precision. To improve user behaviour pattern mining, a two-level clustering process is introduced as Ensemble Fuzzy K-Means with Logit Boost Clustering (EFK-LBC) technique to extract the weblog. In this technique, a preprocessing step is included to remove redundant data and choose reliable log fles. The Fuzzy-K means clustering technique is used to identify behavioural patterns exhibited by recurrent users. Finally, the Logit Boost Clustering method is introduced to the data,that help in generating a strong cluster. Clustering of web users frequent behavioural patterns using the Logit Boost ensemble technique helps the proposed EFK-LBC method to improve newlinethe accuracy up to 88% and reduce the clustering time by 20% compared with existing approaches. Though the proposed EFK-LBC technique performs better for user identifcation, the different initialization of clusters provides various fnal clustering results.
- Source
- Author's Submission
- Date
- 2023-01-01
- Publisher
- Christ(Deemed to be University)
- Subject
- Computer Science and Engineering
- Rights
- Open Access
- Relation
- 61000217
- Format
- Language
- English
- Type
- PhD
- Identifier
- http://hdl.handle.net/10603/479940
Collection
Citation
R, Gokulapriya, “Web User Access Log Analytics Using Neural Learning, Regression and Logit Boost Clustering Techniques for Accurate User Behavioural Pattern Identification,” CHRIST (Deemed To Be University) Institutional Repository, accessed February 23, 2025, https://archives.christuniversity.in/items/show/12270.