20.500.12556/DKUM-82743 Obdelava in gručenje naravnih besedil v programskem okolju python diplomsko delo Natural Language Processing and Clustering in Python Strojno učenje, ki se ukvarja z obravnavo naravnega jezika, je vseprisotno v našem vsakdanjiku. V tem diplomskem delu smo si podrobno pogledali kaj je naravni jezik in kaj zajema njegova obravnava. Opisali smo knjižnice za delo z njim in jih primerjali. Prav tako pa smo opredelili tip nenadzorovanega učenja – gručenja, ki se pogosto uporablja nad naravnimi besedili. V okviru dela smo opisali postopek gručenja in nekaj algoritmov. Nad scenarijem celotne serije Igra prestolov smo izvedli sentimentalno analizo stavkov in opravili gručenje nad podatki. Narisani grafi, so nam pokazali zanimiv rezultate gručenja po sentimentih. Gruče smo kontekstualizirali s pomočjo poznavanja vsebine serije in dogajanja v njej. Machine learning that processes natural language is a part of our everyday lives. In this thesis, we researched what natural language is and what its processing covers. We described the libraries for working with it and compared them. We also defined a type of unsupervised learning - clustering, which is often used when dealing with natural language. As part of the work, we described the clustering process and some algorithms. We performed a sentimental sentence analysis on the script of the entire series Game of Thrones and performed clustering on the data. The drawn graphs showed us interesting results of clustering by sentiments. We contextualized the clusters with the help of knowing the content of the series and the events that take place in it. gručenje obdelava naravnega jezika predprocesiranje besedila knjižnice za obdelavo naravnega jezika metoda voditeljev clustering natural language processing text preprocessing natural language processing libraries Kmeans algorithm true true false [E. Štingl] Slovenski jezik Angleški jezik Diplomsko delo/naloga 2022-08-31 12:45:43 2022-10-24 13:50:19 2023-01-12 03:10:32 0000-00-00 00:00:00 2022 0 Maribor 0 Maribor 1 spletni vir (1 datoteka PDF ([VIII], 28 f.)) 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 004.85:004.93'14(043.2) 137176579 UN_Stingl_Eva_2022.pdf UN_Stingl_Eva_2022.pdf 1 4D3B152009994BABF1E19CC204379EC3 dfbc786e8b89a11f6d717e2e4d661815ecf5cc00638e432bc88118c89af85d7b 31ce01f4-291d-11ed-b1a3-00155d000105 https://dk.um.si/Dokument.php?lang=slv&id=161368 Fakulteta za elektrotehniko, računalništvo in informatiko 0 0 0