| Abstrakt: |
The paper proposes an adaptive filtering model based on website data. It discusses three approaches to data filtering: content-based filtering, collaborative filtering, and hybrid filtering. These approaches aim to tailor the presentation of recommended links to users based on their needs. In addition, various methods of data filtering and processing are studied. Emphasis is placed on adding filtering to the user experience to personalize content. In several person-oriented methods studied during the research, it is proposed to identify pages that have not been seen by the end user, but which satisfy the constant needs, taking into account the relevant search profile. In the study, a comparative analysis of clustering algorithms was conducted, the most suitable algorithm was found to be the CLOPE algorithm. The cases that should be recognized by the personalization system that takes into account the current search query are highlighted. Based on the clustering of search and navigation profiles, a new personalization model, method and algorithm was developed, taking into account the constant and current needs of the end user satisfaction. [ABSTRACT FROM AUTHOR] |
