Optimizacija performansi n-grama u sustavu za strojnu provjeru pravopisa Hascheck

master's thesis

• Title (english): N-Gram Performance Optimization in Hascheck Spelling Checker
• Author: Tomislav Bradarić
• Mentor: Bruno Blašković (mentor)
• Committee member: Bruno Blašković (predsjednik povjerenstva)
• Committee member: Šandor Dembitz (član povjerenstva)
• Committee member: Gordan Gledec (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2016-07-08, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Razvoj i održavanje programske potpore za programske alate složen je proces u kojemu je od iznimnog značenja kontinuirano poboljšanje performansi sustava. Ovaj rad analizira prednosti, mane i performanse jezgrenih funkcionalnosti postojećeg podsustava za podršku spremanju i analizi n-grama u strojnom pravopisnom provjerniku Hascheck. Nakon toga, razrađuje ciljeve optimizacije pozivajući se pritom na relevantne pristupe sličnim problemima. Gradeći na postojećim tehnologijama za upravljanje podatcima, ostvarena je implementacija novog n-gramskog podsustava koji se oslanja na dva pristupa upravljanju podatcima. Rad optimiziranog sustava je potom ilustriran na odabranom skupu primjera, izmjerene su njegove performanse, te je donesen zaključak o iskoristivosti njegovih pristupa problemu.

Abstract (english)

Development and maintenance of software is a complex process in which constant improvement of performance is of exceptional importance. This thesis performs a pain point and performance analysis of the existing subsystem for the analysis and storage of n-grams which is used by the automated spellchecker Hascheck. Following that, the goals of the optimization are defined in accordance with relevant problems and solutions. Building upon existing data technologies, a new n-gram subsystem which relies upon two approaches to data management is designed and implemented. The functionality of the optimized system is illustrated on a chosen set of examples, its performance is measured, and a conclusion that relates to the usability of its approaches to the problem is brought.

• Keywords: baze podataka; SQL; NoSQL; skalabilnost; pretraga teksta; pravopisni provjernik; nadzirano učenje; PostgreSQL; ElasticSearch
• Keywords (english): databases; SQL; NoSQL; scalability; text search; automated spellchecker; supervised learning; PostgreSQL; ElasticSearch
• Language: croatian
• URN:NBN: urn:nbn:hr:168:565137
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2018-02-06 08:00:32