AN ANALYTICAL REVIEW OF CONTENT-BASED AND COLLABORATIVE FILTERING IN RECOMMENDER SYSTEMS
Advanced Information Technology · 2025 · (1 (4)) · 54-66
DOI: 10.17721/ait.2025.1.07 ↗Kurz gefasst
Fachartikel in der Zeitschrift Advanced Information Technology, 2025. Verfasst von Maksym SHEVCHENKO. Die Arbeit nennt 38 Literaturangaben.
Abstract (Originalsprache)
Background
With the rapid growth of digital content, recommender systems are becoming a key tool for providing personalized offers. They contribute to the discovery of new movies, music, and products, maintaining user interest in using platforms. The relevance of researching recommender system algorithms is due to the need to improve their work to satisfy individual user preferences. This paper presents a review and analytical study of recommender system algorithms. The purpose of this paper is to systematize, classify, and critically analyze two main approaches in recommender systems: content-based filtering and collaborative filtering.
Methods
A review of existing recommender system methods, a comparative and analytical assessment.
Results
The work analyzes recommender system algorithms. A formal definition of the recommendation problem is given, where user preferences are modeled as a functional dependence on object properties. Within the framework of content-based filtering, the use of classification algorithms, such as a Naive Bayes classifier and decision trees, as well as the Rocchio algorithm, which uses relevant feedback to update the user profile, is considered. The strengths and weaknesses of different similarity measures between vectors are analyzed. In collaborative filtering, the memory-based approach (user-based and item-based methods) and model-based techniques with an emphasis on the k-NN algorithm are investigated. To overcome the shortcomings of individual methods, a hybrid approach is proposed that combines their advantages. Methods for integrating systems into a hybrid model are presented, which allows improving the accuracy of recommendations.
Conclusions
The results of the work highlight the features of the specified filtering methods, demonstrate the impact of the implementation of algorithms and input data on the accuracy of recommendations and response time. The analysis of shortcomings emphasizes the importance of the combined use of filtering algorithms to improve the efficiency of recommendation systems, which makes the hybrid approach a promising direction for further research and implementation.
Zitieren
APA
SHEVCHENKO, M. (2025). AN ANALYTICAL REVIEW OF CONTENT-BASED AND COLLABORATIVE FILTERING IN RECOMMENDER SYSTEMS. Advanced Information Technology, 54-66. https://doi.org/10.17721/ait.2025.1.07
Vancouver
SHEVCHENKO M. AN ANALYTICAL REVIEW OF CONTENT-BASED AND COLLABORATIVE FILTERING IN RECOMMENDER SYSTEMS. Advanced Information Technology. 2025:54-66. doi:10.17721/ait.2025.1.07
BibTeX
@article{SHEVCHENKO2025,
author = {SHEVCHENKO, Maksym},
title = {AN ANALYTICAL REVIEW OF CONTENT-BASED AND COLLABORATIVE FILTERING IN RECOMMENDER SYSTEMS},
journal = {Advanced Information Technology},
year = {2025},
number = {1 (4)},
pages = {54--66},
doi = {10.17721/ait.2025.1.07}
}
EUVIMED