Sustav preporučivanja knjiga zasnovan na suradničkom filtriranju
| Autor: | Majača, Josip |
|---|---|
| Přispěvatelé: | Delač, Goran |
| Jazyk: | chorvatština |
| Rok vydání: | 2020 |
| Předmět: |
item-item algoritam
recommender system user-user algorithm matrica sličnosti TEHNIČKE ZNANOSTI. Računarstvo sustav preporučivanja suradničko filtriranje kosinusna sličnost user-user algoritam item-item algorithm TECHNICAL SCIENCES. Computing collaborative filtering cosine similarity similarity matrix |
| Popis: | Ovaj rad se bavi sustavima za preporučivanje zasnovanim na suradničkom filtriranju, specifičnije dvama algoritmima takvih sustava: algoritmu zasnovanom na korisnicima i algoritamu zasnovanom na predmetima. Detaljno se objašnjava svaki od algoritama zajedno s prednostima i nedostacima. Opisan je dio programskog ostvarenja algoritama sustava preproučavanja knjiga u programskom jeziku Python. Na kraju je provedena validacija rezultata algoritama kako bi se usporedila pogreška algoritama te su na osnovu toga izvedeni zaključci. This thesis deals with recommender system based on collaborative filtering, more specifically two algorithms: a user-based algorithm and an item-based algorithm. Each of the algorithms is explained in detail with its own advantages and disadvantages. A part of Python implementation of the book recommender system is described. Finally, validation of the algorithm result was performed to compare the algorithms errors and conclusions were drawn on that basis. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|