Impact of the Context Relevancy on Ratings Prediction in a Movie-Recommender System

Abstract

Original scientific paperRecommender systems are a popular and a highly researched way of helping users get to their desired content in the huge amount of available data, and services online. Understanding the situation in which users consume the items was shown to improve the recommendation process. For that reason, context-aware recommender system (CARS) employs contextual information in order to enhance the user's model and to improve the recommendations. An issue that is still open is how to decide which pieces of contextual information to acquire and how to incorporate them into CARS, since using irrelevant piece of contextual information could have a negative impact on the recommendations. We propose a methodology for detecting which pieces of contextual information contribute to explaining the variance in the ratings, based on statistical testing. We also inspect the impact of the detected relevant pieces of contextual information on the ratings prediction based on the matrix-factorization algorithm. The experiment was conducted on the MovieAT database. The results showed a significant difference in the ratings prediction using the relevant and the irrelevant pieces of contextual information. We also confirmed the positive impact of the relevant, and negative impact of the irrelevant pieces of contextual information with respect to the uncontextualized model. Key words: Personalization, Recommender systems, Context-awarenessUtjecaj relevantnosti konteksta na predvi anje ocjena u sustavu za preporuke filmova. Sustavi za preporuke (eng. recommender systems) predstavljajučest i vrlo istražen način pružanja pomoći korisnicima u svrhu pronalaska željenog sadržaja u velikoj količini dostupnih podataka i usluga. Pokazalo se da uvid u situaciju u kojoj korisnici koriste sadržaj doprinosi kvaliteti preporuka. Zbog toga, konteksta svjesni sustavi za preporuke (eng. context-aware recommender systems CARS) koriste kontekstne informacije kako bi poboljšali model korisnika i time kvalitetu preporuka. Jedan od neriješenih problema je kako odlučiti koje kontekstne informacije je potrebno sakupiti i kako ih upotrijebiti u CARSu, budući da upotreba nebitnih kontekstnih informacija može imati negativan utjecaj na kvalitetu preporuka. Mi predlažemo metodologiju za otkrivanje onih kontekstih informacija koje doprinose objašnjavanju varijabilnosti ocjena za sadržaje, utemeljenu na statističkom testiranju. Tako er, istražujemo utjecaj otkrivenog bitnog konteksta na predvi anje ocjena utemeljeno na algoritmu faktorizacije matrica. Eksperiment je proveden na bazi podataka MovieAT. Rezultati su pokazali znatnu razliku u predvi anju ocjena prilikom korištenja bitnog i nebitnog konteksta. Ujedno je potvr en i pozitivan utjecaj bitnog, odnosno negativan utjecaj nebitnog konteksta, u odnosu na sustav koji ne koristi kontekst, što upućuje na važnost i kvalitetu detekcije.

Tourist Recommender Systems Based on Emotion Recognition—A Scientometric Review

“…Most of the studies used various stimuli to provoke the emotional states of the participants. Various methods have been described, including viewing video clips [11,163,165,173,[182][183][184], images [27,49,169], listen to music [8,12,185,186], read texts [6,32], and doing physical activities [177,187,188]. Emotions can be assessed through subjective and objective methods.…”

Time-Bin-Based Neighbourhood Algorithm for Temporal Effects in Recommendation Systems

“…Collaborative filtering (CF) is a commonly used RS implementation technique in the literature [1][2][3]. From the application perspective, movie-based RS is a favourite research subject in CF [4][5][6]. Since the development of scientific movie datasets, such as MovieLens [7,8] and Netflix Prize [9], more studies are investigating this research topic.…”

An Algorithm for Density Enrichment of Sparse Collaborative Filtering Datasets Using Robust Predictions as Derived Ratings

“…Finally, it is worth noting that the CF DR algorithm, proposed in this paper, can be combined with other works in the domain of CF, targeting either to (1) enhance rating prediction accuracy [14][15][16][17], (2) improve recommendation quality [18][19][20][21], (3) increase rating prediction computation efficiency [22][23][24][25], or (4) further increase prediction coverage in CF-based RSs [26][27][28][29].…”