20.500.12556/DKUM-87011 Primerjava algoritmov za analizo sentimenta v filmskih kritikah magistrsko delo Sentiment analysis on movie reviews: a comparative study of algorithms V magistrskem delu smo primerjali različne pristope za klasifikacijo sentimenta kritik filmov. Naš cilj je bil doseči čim višjo točnost pri klasifikaciji kritik. Uporabili smo algoritme, kot so metoda podpornih vektorjev, naključni gozdovi, naivni Bayes, odločitvena drevesa, k-najbližjih sosedov, logistična regresija in glasovanje. Rezultate smo merili z metriko točnosti. Ugotovili smo, da večji n-grami in algoritmi glasovanja dosegajo najboljšo točnost. Za najhitrejšo in najbolj točno klasifikacijo priporočamo algoritem glasovanja brez metode podpornih vektorjev, saj je bila ta najpočasnejša. In the master’s thesis, we compared different approaches for sentiment classification of movie reviews. Our goal was to achieve the highest possible accuracy in classifying reviews. We used algorithms such as support vector machines, random forests, naive Bayes, decision trees, k-nearest neighbors, logistic regression and voting. The results were measured using the accuracy metric. We found that larger n-grams and voting algorithms achieve the best accuracy. For the fastest and most accurate classification, we recommend the voting algorithm without support vector machines, as the latter was the slowest of all algorithms. analiza sentimenta n-grami algoritmi za klasifikacijo točnost sentiment analysis n-gram classification algorithms accuracy true true false [V. Milutinović] Slovenski jezik Angleški jezik Magistrsko delo/naloga 2024-02-06 19:06:18 2024-03-01 12:36:42 2024-04-06 04:17:34 0000-00-00 00:00:00 2024 0 Maribor 0 Maribor 1 spletni vir (1 datoteka PDF (VII,65 f.)) 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 004.93.021(043.2) 191510275 MAG_Milutinovic_Virdzinija_2024.pdf MAG_Milutinovic_Virdzinija_2024.pdf 1 A8A73D9F6559B3FDCCB10C80422D33B5 516999d14ad0921cb94320c338d32625a9f462a395d6165170a69d47fee28904 1b4f69e2-c522-11ee-b2aa-00155d000105 https://dk.um.si/Dokument.php?lang=slv&id=175971 Fakulteta za elektrotehniko, računalništvo in informatiko 0 0 0