Open AccessUse of Kemeny median in the algorithm of forming recommendation
Author(s) -
Oksana S. Zhurakovska,
Illa Yu. Kochubey
Publication year - 2020
Publication title -
sistemnì doslìdžennâ ta ìnformacìjnì tehnologìï
Language(s) - English
Resource type - Journals
eISSN - 2308-8893
pISSN - 1681-6048
DOI - 10.20535/srit.2308-8893.2020.4.05
Subject(s) - ranking (information retrieval) , computer science , recommender system , set (abstract data type) , object (grammar) , algorithm , cluster (spacecraft) , information retrieval , data mining , theoretical computer science , artificial intelligence , programming language
The relevant nowadays question of development of the algorithmic support of recommender systems is considered. The article is devoted to the solution of the problem of forming recommendations to new users, which is based on the ideas of transition from the matrix "user-object" to the ranking of objects and the formation of recommendations to the user of the active cluster based on the construction of the resulting ranking, which is a Kemeny median on a set of rankings. The choice of Kemeny median as the resulting ranking and the choice of algorithm for its construction are justified. To reduce the complexity of calculations, it is suggested to perform aggregation of information and to use it in forming of ranking recommendations, which are based on a set of "generalized experts" for this cluster. The efficiency of the developed algorithmic support was studied and the results and recommendations were given.