Toxic comment classification application in Python

Vladimir Zinovev

Bakalářská práce

Abstract:

In this thesis the development of natural language processing models is presented.The models are coded to be able to recognize the different possible connotations and thesentiment of a random analyzed text. The program will process given string values andmake reasonable predictions of its sentiment polarity. The predictions are based on thestring content.For this purpose, I would like to use Python tools and their advantage in the fieldsof big size data processing and the possibility to create a working neural network fromscratch for some scientific-academics non-commercial computing or for personal non-commercial use. So this programming language is very convenient for this project taskassignment as it provides the developer with a good toolkit for artificial intelligencemodeling in many senses. Also, the wide range of open-source natural language processinglibraries for Python available.Natural language processing is a huge area of machine learning studies that couldbe used in many different ways such as education, business, finance, robotics, biology, etc.Hence it could be useful and pertinent as for learning and research intents so forentrepreneurial or business objectives. Modern neural networks are widely implemented tofulfill many needs and productively process the big amounts of digital data.

Keywords

data science, machine learning, big data, Python, artificial intelligence, text analysis, strojové učení, umělá inteligence, textová analýza

Jazyk práce: angličtina

Datum vytvoření / odevzdání či podání práce: 23. 3. 2020

Obhajoba závěrečné práce

• Vedoucí: Ing. Jiří Brožek, Ph.D.
• Oponent: Petr Hanzlík, Ph.D.

Česká zemědělská univerzita v Praze

Provozně ekonomická fakulta

Bakalářský studijní program / obor:
Systems Engineering and Informatics / Informatics