diplomsko delo
Domen Dobnikar (Author), Aleš Smrdel (Mentor)

Abstract

V primeru aplikacij, ki se ukvarjajo s prodajo, ponudbo izdelkov ali pa kakšnimi drugačnimi vsebinami, želimo uporabniku zagotoviti čim boljšo izkušnjo. Eden izmed že zelo dolgo prisotnih načinov je z uporabo priporočilnih sistemov, pri katerem uporabniku ponujamo predvsem tisto vsebino, za katero predvidevamo, da bo za takega uporabnika zanimiva glede na informacije, ki jih imamo v zvezi z uporabnikom, in mu s tem prihranimo čas iskanj, hkrati pa povečamo možnost nakupa. V diplomski nalogi je predstavljen pregled nekaj izbranih priporočilnih sistemov, opisan postopek izdelave in implementacije takih sistemov v spletno trgovino ter njihovo testiranje s strani razvijalca in nekaj nakjučnih uporabnikov. Priporočilnih sistemov je več vrst. Vsak sistem ima svoje prednosti in slabosti. V okviru diplome sem jih raziskal in opisal kar se da natančo in poskušal ugotoviti, kateri delujejo najbolje v primeru spletne trgovine. Implementirane sisteme sem testiral po dveh kriterijih: kvaliteta priporočil in odzivnost sistema. Za najboljša sta se izkazala sistem \emph{Najbolj priljubljen} in sistem \emph{Komplementarni izdelki}. Za razvoj spletne trgovine in implementacijo priporočilnih sistemov sem uporabil programski jezik Python, spletni okvir za spletno programiranje Django, knjižnico za strojno učenje Scikit-learn in pa knjižnico za matematične funkcije ter n-dimenzionalne tabele Numpy.

Keywords

spletna trgovina;priporočilni sistemi;testiranje;visokošolski strokovni študij;diplomske naloge;

Data

Language: Slovenian
Year of publishing:
Typology: 2.11 - Undergraduate Thesis
Organization: UL FRI - Faculty of Computer and Information Science
Publisher: [D. Dobnikar]
UDC: 004.774(043.2)
COBISS: 98179075 Link will open in a new window
Views: 112
Downloads: 28
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Review, implementation and testing of various recommendation systems for online shopping
Secondary abstract: In case of applications, that are dealing with selling, offering products or some other content, we want to provide users with good user experience. One of the long present ways, isby using a recommendation system, where we can recommend content, for which we think would be interesting to the user, based on the information we have regarding that user. With this, we can save the user time for searching, and at the same time improve possibility of purchase. Diploma thesis presents an overview of some selected recommendation systems, the process of making and implementing such systems into online store and their testing by the developers and random users. There are several types of recommendation systems. Each has its pros and cons. In the scope of the diploma I researched and described as exactly as possible, which perform well in case of online store. I tested the implemented systems according to two criteria: recommendation quality and responsiveness of the system. The best systems proved to be The most popular system and Complementary products. To develop the online store and implement the recommendation system I used Python programming language, web framework for web programming Django, library for machine learning Scikit-learn and library for mathematical functions and n-dimensional tables Numpy.
Secondary keywords: online store;recommendation systems;testing;computer science;computer and information science;diploma;Spletne aplikacije;Računalništvo;Univerzitetna in visokošolska dela;
Type (COBISS): Bachelor thesis/paper
Study programme: 1000470
Embargo end date (OpenAIRE): 1970-01-01
Thesis comment: Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages: 40 str.
ID: 14545335