Games Recommender System Based on Collaborative Filtering

Sustavi preporučivanja (engl. recommender systems) softverska su rješenja i tehnike čija je svrha preporučiti sadržaj korisniku. Sustavi zasnovani na suradničkom filtriranju (engl. collaborative filtering) odluku o preporuci predmeta korisniku temelje na njegovu prošlom ponašanju kao i ponašanju drugih korisnika. Vrednovanje sustava provodi se s ciljem utvrđivanja koliko dobro radi sustav preporučivanja. U sklopu rada predstavljena je implementacija web pauka (engl. web spider) za prikupljanje podataka o ocjenama igara s web stranice Metacritic. Implementiran je sustav preporučivanja zasnovan na suradničkom filtriranju i provedeno je njegovo vrednovanje.Recommender systems are software tools and techniques that provide recommendations for content to users. Recommender systems based on collaborative filtering build their recommendations having taken into consideration the past behaviors of both the target user and of users similar to the target user. System validation is used to determine how well the recommender system works. As part of this thesis, a web spider implementation for web scraping Metacritic’s website is presented. A recommender system based on collaborative filtering is implemented and validated