DOI

In this paper we propose a method of movie series recommender system development. Our recommender system is content-based, and movie series are represented by their scripts. We experiment with several semantic similarity measures, lexico-morphological metrics, keywords and vector space models to extract similar movie series. Evaluation is conducted in the experiment with informants. The best results are achieved by distributional semantic approach (i.e., using word2vec technology).
Translated title of the contributionМеры и модели определения сходства текстов для построения рекомендательной системы
Original languageEnglish
Title of host publicationInternet Science - 5th International Conference, INSCI 2018, St. Petersburg, Russia, October 24-26, 2018, Proceedings
PublisherSpringer Nature
Pages181-193
Number of pages12
Volume11193 LNCS
ISBN (Electronic)978-3-030-01437-7
ISBN (Print)978-3-030-01436-0
DOIs
StatePublished - 25 Sep 2018

Publication series

NameLecture Notes in Computer Science
PublisherSpringer
Volume11193
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

    Research areas

  • Recommender system, Clustering algorithm, Movie series scripts, Vector space model, Semantic similarity measure

    Scopus subject areas

  • Computer Science (miscellaneous)

ID: 37544184