Sentimentu analīze tekstiem krievu valodā
Loading...
Date
Authors
Advisor
Journal Title
Journal ISSN
Volume Title
Publisher
Latvijas Universitāte
Language
N/A
Abstract
Kvalifikācijas darbs tika veidots ar mērķi analizēt „rus.tvnet.lv” portāla komentārus krievu valodā, jo latviešu valodai līdzīgs projekts jau ir izveidots [1]. Programmas izstrādes sākumā bija pieņemts izstrādāt sistēmu, kas spētu izgūt portāla komentārus un ar mašīnmācīšanās algoritmu palīdzību noteikt negatīvus un pazemojošus komentārus. Projekta gaitā tika izlemts paplašināt programmas darbību, ļaujot lietotājiem pašiem noteikt kāda veida komentāri tiks atlasīti.
Par mašīnmācīšanās algoritmu tika izvēlēts naivais Baiesa klasifikators.
Tika izmantotas PHP un JavaScript programmēsanas valodas, kā arī AJAX tehnoloģija.
Sistēma sastāv no 2 moduļiem: tīmekļa vietnes un „Chrome” pārlūkprogrammas spraudņa.
The qualification work was created to analyze comments of „rus.tvnet.lv” portal that were written in Russian language because a similar project for Latvian language already exists [1]. In the beginning of the development process the task was to develop a system that could obtain comments from the portal and use machine learning algorithms to detect negative and humiliating comments. During the development process it was decided to expand the scope of activity of the program by letting users decide themselves what type of comments would be selected. Naïve Bayes classifier was chosen as the machine learning algorithm. PHP and JavaScript programming languages and AJAX technology were used. The system consists of two modules: website and “Chrome” browser extension.
The qualification work was created to analyze comments of „rus.tvnet.lv” portal that were written in Russian language because a similar project for Latvian language already exists [1]. In the beginning of the development process the task was to develop a system that could obtain comments from the portal and use machine learning algorithms to detect negative and humiliating comments. During the development process it was decided to expand the scope of activity of the program by letting users decide themselves what type of comments would be selected. Naïve Bayes classifier was chosen as the machine learning algorithm. PHP and JavaScript programming languages and AJAX technology were used. The system consists of two modules: website and “Chrome” browser extension.