Nlp- та LLM-орієнтована система персоналізованих рекомендацій фільмів

Abstract

This text discusses the development of a recommendation system for movies based on text analysis, aiming to simplify the process of finding suitable films. The system aggregates reviews from reliable sources like Metacritic and Rotten Tomatoes and uses user profile data to suggest movies, especially during the «cold start» phase. It also customizes recommendations based on user preferences and allows for joint recommendations for couples by analyzing both users’ tastes.