Options
ქართული ენის ამომცნობი სისტემა
Date Issued
2019
Author(s)
Advisor(s)
Institution
Abstract
In this thesis has been examined a system which implements recognition and processing of the Georgian language from various information sources, for example, such as for all well known Wikipedia, Facebook and other numerous social and information networks, but for recognition and processing are not used the algorithms of standard stemming and lemmatization, as the Georgian language and its peculiarities make use of the standard algorithms impossible, that is why a new stemming algorithm became necessary which would implement the processing of the text and which in this concrete case is based exclusively on the base of the initial forms of conjunctions, pronouns and nouns.
It should also be mentioned that similar systems are not so many for the Georgian language and in this thesis the positive sides of all existing systems, one way or another, are stipulated and their negative sides are improved. Exactly thus and so, this system gave us certain results, these results give us the basis of the positive outcome and that is why we can say with certainty that this system can be an initial stage of commencement, development and creation of a full-fledged system that will implement complete recognition and processing of the Georgian language.
It should also be mentioned that similar systems are not so many for the Georgian language and in this thesis the positive sides of all existing systems, one way or another, are stipulated and their negative sides are improved. Exactly thus and so, this system gave us certain results, these results give us the basis of the positive outcome and that is why we can say with certainty that this system can be an initial stage of commencement, development and creation of a full-fledged system that will implement complete recognition and processing of the Georgian language.
Subjects
File(s)
Loading...
Name
MA Thesis Kobaxidze Davit.pdf
Description
ქართული ენის ამომცნობი სისტემა
Size
1.12 MB
Format
Adobe PDF
Checksum
(MD5):674c7c71391c8a3d2b3b4efd857a8d90