DOI:https://doi.org/10.15407/kvt211.01.040
Cybernetics and Computer Engineering, 2023, 1(211)
ZOSIMOV V.V., DSc (Engineering), Associate Professor,
Professor of the Department of Applied Information Systems,
https://orcid.org/0000-0003-0824-4168,
e-mail: zosimovv@gmail.com
Taras Shevchenko National University of Kyiv,
60, Volodymyrska st., Kiyv, 01033, Ukraine
PROBABALISTIC APPROACH TO RANKING SEARCH RESULTS USING BAYESIAN BELIEF NETWORKS
Introduction. This paper proposes a probabilistic approach to ranking search results using Bayesian Belief Networks (BBN). The proposed approach utilizes BBN to model the relationships between search queries, web pages, and user feedback, and to calculate the probability of a web page being relevant to a specific query. The approach takes into account various factors, such as keywords, page relevance, domain authority, and user feedback to generate a ranking score for each search result.
The purpose of the article is to conduct an analysis on the feasibility of creating a search engine that uses BBNs and probabilistic ranking methods for improving the accuracy and efficiency of search results.
Results. The proposed approach was evaluated on a real-world dataset, and the results showed its effectiveness. Overall, the results suggest that the use of BBNs can provide a promising approach to enhancing search engine performance and user experience. The approach’s effectiveness is attributed to its ability to model and reason about uncertainty and dependencies among variables, and its consideration of various factors, such as keywords, page relevance, domain authority, and user feedback.
Conclusions. The proposed method has the potential to improve search relevance, reduce user frustration, and increase user satisfaction. However, further research is needed to optimize the proposed approach and to explore its applicability in different contexts. Overall, the study suggests that BBNs can provide a valuable tool for developing more effective and user-friendly search engines. Moreover, the use of Sphinx as a base search system shows promise in enabling the proposed approach to be integrated into practical search systems. Nonetheless, further research is needed to optimize the approach and evaluate its applicability in different contexts.
Keywords: search engine, ranking, Bayesian Belief Networks, probabilistic model, information retrieval.
REFERENCES
1 Baeza-Yates, R., & Ribeiro-Neto, B. Modern Information Retrieval: The Concepts and Technology behind Search. Addison-Wesley Professional. 2011.
2 Sattari P. Bayesian deep reinforcement learning: A survey. Journal of Machine Learning Research. JMLR.org. 2020, Vol. 21, pp. 1-35.
3 Agichtein, E., Brill, E., & Dumais, S. Improving Web Search Ranking: Beyond the Query-Document Similarity. Synthesis Lectures on Information Concepts, Retrieval, and Services. 2006, Vol. 1(1), pp. 1-136.
4 Chau M. Spidering and Filtering Web Pages for Vertical Search Engines. Proceedings of The Americas Conference on Information Systems. AMCIS 2002 Doctoral Consortium, Dallas, TX, USA, 2002.
5 Zosimov V.V., Bulgakova O.S., Pozdeev V.O. Complex internet data management system. Advances in Intelligent Systems and Computing. AISC. 2021, Vol.1246, pp. 639-652.
https://doi.org/10.1007/978-3-030-54215-3_41
6 Pelt M. Uncertainty quantification in deep learning using Bayesian convolutional neural networks. Journal of Computer Vision. 2019, Vol. 126, pp. 617-635.
7 Zosimov. V.V., Bulgakova. O.S. Calculation the Measure of Expert Opinions Consistency Based on Social Profile Using Inductive Algorithms. Advances in Intelligent Systems and Computing. 2020. Vol. 1020. pp. 622-636.
https://doi.org/10.1007/978-3-030-26474-1_43
8 Bendersky, M., Croft, W. B., & Zhang, J. Predicting query performance via classification. Proceedings of the ACM Conference on Information and Knowledge Management (CIKM). 2010, pp. 79-88.
9 Hron J. Probabilistic programming for deep learning: A review. Machine Learning Research. 2018, Vol. 19, pp 1-41.
10 Gallego C. A review of Bayesian deep learning techniques and their application to computer vision problems. Big Data Analytics, IGI Global. 2018, pp. 11-25.
11 Guo C. Deep Bayesian active learning for neural networks. Journal of Machine Learning Research, JMLR.org. 2017, Vol. 18, pp. 1-47.
12 Sattari P. Bayesian deep reinforcement learning: A survey. Journal of Machine Learning Research, JMLR.org. 2020, Vol. 21, pp. 1-35.
13 Nalisnick M. Deep Bayesian neural networks with many irrelevant inputs. Proceedings of the 35th International Conference on Machine Learning. 2019, Vol. 97, pp. 1748-1757.
14 Official Sphinx search system site. URL: Sphinx http://www.sphinxsearch.com/
Received 23.01.2023