Analýza vlivu reprezentace textu na kvalitu klasifikace pomocí metod umělé inteligence

Mikhail Larkin

Master's thesis

Analysis of Text Representation Influence on the Quality of Classification by Artificial Intelligence Methods

Abstract:

Tato práce se zabývá analýzou vlivu metod reprezentace textu na kvalitu kvalifikace pomocí metod umělé inteligence. Cílem práce je porovnat populární metody reprezentace textu a posoudit jejich vliv na kvalitu klasifikace textu pomocí technik strojového a hlubokého učení. Byl analyzován vliv metod reprezentace textu, jako jsou Bag of Words, TF-IDF, Word2Vec, GloVe, FastText a Doc2Vec. Jako klasifikátory byly zvoleny Random Forest, Support Vector Machine, Multilayer Perceptron a Convolutional Neural Network.

Abstract:

This thesis deals with the analysis of the influence of text representation methods on the quality of qualification using artificial intelligence methods. The aim of this paper is to compare popular text representation methods and assess their impact on text classification quality using machine and deep learning techniques. The influence of text representation methods such as Bag of Words, TF-IDF, Word2Vec, GloVe, FastText and Doc2Vec has been analyzed. Random Forest, Support Vector Machine, Multilayer Perceptron and Convolutional Neural Network were chosen as classifiers

Keywords

Klasifikace textu, NLP, BoW, TF-IDF, Word2Vec, GloVe, FastText, Doc2Vec, CNN, MLP, Random Forest, SVM

Language used: Czech

Date on which the thesis was submitted / produced: 25. 5. 2023

Thesis defence

Tomas Bata University in Zlín

Faculty of Applied Informatics

Master programme / field:
Information Technologies / Software Engineering