Movie Recommender System Based on Collaborative Filtering

Ovaj rad obrađuje sustave za preporučivanje. Navedene i objašnjene su tri metode sustava za preporučivanje, preporučivanje zasnovano na sadržaju, suradničko filtriranje i hibridno preporučivanje. Detaljnije je teorijski objašnjena metoda suradničkog filtriranja koja koristi dva pristupa, preporučivanje zasnovano na sličnosti korisnika i preporučivanje zanovano na sličnosti proizvoda. Programska izvedba suradničkog filtriranja ostvarena je u programskom jeziku Python, objašnjena je uz priložene dijelove koda te su prikazani rezultati preporučivanja i validacije sustava.This thesis discusses recommendation systems. Three recommendation system methods are mentioned and explained, content based recommendation, collaborative filtering and hybrid method of recommendation. Collaborative filtering method is more detailed explained with two methods, model based and memory based approach. Program solution of collaborative filtering based on memory is accomplishe in Python language, it is explained with enclosed parts of code and system results of recommendation and validation are presented