Popravljanje vejic v slovenskih besedilih z orodjem LanguageTool

Piškur, Karin (author), Robnik Šikonja, Marko (mentor)

Cilj diplomskega dela je dodati pravila za pisanje vejice v orodje LanguageTool. S pomočjo analize korpusa Lektor smo raziskali, katera pravila za postavljanje vejic v slovenščini piscem povzročajo največ težav. Glede na dobljene rezultate smo v Slovenskem pravopisu 2001 podrobneje raziskali pravila za pisanje vejice pred vezniki »in«, »ali« ter »da«. Po končani raziskavi smo poskusili implementirati čim več pravil za odprtokodno orodje LanguageTool, ki ga lahko uporabljamo kot samostojen program, spletni vmesnik ali v odprtokodnih pisarniških paketih LibreOffice in OpenOffice. Nekaj pravil smo uspešno implementirali, vendar ne vseh, saj bi za to potrebovali označevalnik besedila, ki ga orodje LanguageTool za slovenski jezik še ne vsebuje. Na koncu smo opravili še evalvacijo pravil, s katero smo preverili njihovo pravilnost, uporabnost in uporabniško izkušnjo.

• Language: Slovenian
• Keywords: LanguageTool, slovenski jezik, vejica, ločila, korpus Lektor
• Work type: Bachelor thesis/paper
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2015
• PID: 20.500.12556/RUL-72287
• Publication date in RUL: 10.09.2015

Secondary language

• Language: English
• Title: Correcting comma placement in Slovene language with LanguageTool

Abstract

The aim of the thesis is to add the rules for comma usage to the LanguageTool program. Using the Lektor corpus, we examined which rules for comma usage are causing the most issues in written Slovene. In view of these results, we analyzed the rules for comma usage before conjunctions »and«, »or« and »that« in Slovenian ortography 2001. After finishing the analysis, we tried to implement comma placement rules for the open source program LanguageTool, which can be used as a stand-alone desktop application, as web interface or in open source office suites LibreOffice and OpenOffice. Some of the rules were successfully implemented. For all of the rules to be implemented, we would need the part-of-speech tagger, which is not a part of the LanguageTool for Slovene, yet. We evaluated the rules, taking their accuracy, applicability and user experience into account.

• Keywords: LanguageTool, Slovenian language, comma, punctuation mark, Lektor corpus