Automatsko označavanje i analiza teksta za LARSP hrvatskog jezika

master's thesis

• Title (english): Automatic text markup and analysis for LARSP of Croatian language
• Author: Bruno Tot
• Mentor: Nikolaj Lazić (mentor)
• Mentor: Vesna Mildner (mentor)
• Committee member: Diana Tomić (predsjednik povjerenstva)
• Committee member: Vesna Mildner (član povjerenstva)
• Committee member: Nikolaj Lazić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Humanities and Social Sciences; (Department of information and Communication sciences); Zagreb
• Granter: University of Zagreb; Faculty of Humanities and Social Sciences; (Department of Phonetics); Zagreb
• Defense date and country: 2019-11-06, Croatia
• Scientific / art field, discipline and subdiscipline: SOCIAL SCIENCES; Information and Communication Sciences
• Scientific / art field, discipline and subdiscipline: HUMANISTIC SCIENCES; Philology; Phonetics

Abstract

Govor je jedan od najznačajnijih načina ljudske komunikacije. Govor i jezik se zajedno razvijaju od rođenja – prvo slušanjem, a zatim imitacijom. Kao i sve ljudske osobine, podložni su poremećajima i oštećenjima. Neka od njih se mogu rehabilitirati zbog čega je važno što prije ih detektirati. U tu svrhu je razvijen LARSP – lingvistički protokol koji objedinjuje procjenu, rehabilitaciju i probir tj. praćenje kako bi omogućio točnu dijagnozu i uspješnu rehabilitaciju. Temelji se na gramatičkoj analizi transkribiranog dječjeg govora. Ručna analiza većeg korpusa teksta dugotrajna je i podložna ljudskim pogreškama. LARSPTool je alat razvijen kako bi olakšao taj postupak. Automatska predobrada i prikupljanje podataka znatno ubrzavaju analizu i omogućuju analizu mnogo većeg korpusa. LARSPTool također uvodi standard i jednoličnost obrade i prikaza podataka što otvara daljnje mogućnosti za praćenje trendova razvoja govora i jezika kod djece kroz generacije. Trenutna verzija ima integriranu podršku za hrvatski, a kasnije verzije bit će modularne. Također, dostupna je samo za Windows operativne sustave, ali postoji mogućnost razvoja verzija i za druge platforme.

Abstract (english)

Speech is one of the most significant methods of human communication. Speech and language start developing together from birth – first by listening then by imitation. Like all human characteristics it is subject to disorders and defects. Some of them can be remedied which is why it is important to detect them as soon as possible. To that end LARSP was developed – a linguistic protocol which unifies assessment, remediation and screening in order to enable correct diagnosis and successful remediation. It is based on grammatical analysis of transcribed children’s speech. Manual analysis of a larger corpus of text is time-consuming and susceptible 32 to human error. LARSPTool is a tool developed to facilitate that process. Automatic preprocessing and data gathering significantly speed up analysis and enable analysis of a much larger corpus. LARSPTool also introduces a standard and uniformity to both data processing and presentation which opens up further possibilities of tracking children’s speech and language development trends over generations. The current version has integrated support for Croatian but later versions will be modular. Also, it is only available for Windows operating systems but there is a possibility of developing versions for other platforms as well.

• Keywords: LARSP; djeca; govor; automatska analiza; jezik
• Keywords (english): LARSP; children; speech; automatic analysis; language
• Language: croatian
• URN:NBN: urn:nbn:hr:131:378134
• Study programme: Title: Information Sciences; specializations in: Computer Science (Teaching, Research), Informatology; Course: Computer Science (Teaching, Research); Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra informacijskih znanosti (magistar/magistra informacijskih znanosti)
• Study programme: Title: Phonetics (double major); specializations in: Clinical Phonetics, Rhetoric, General Phonetics; Course: General Phonetics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra edukacije fonetike (magistar/magistra edukacije fonetike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2020-01-09 14:19:40