Institutional Repository of Vilnius University
Propagandos ir dezinformacijos aptikimas mašininio mokymo algoritmais /
Abstract
dc:descriptionThe main objective of this thesis is to identify, test and compare publicly available propaganda and disinformation datasets, preprocessing tools and known classification models. This objective and all tasks have been completed successfully. During dataset and machine learning classifier investigation "Fake news" and "LIAR" datasets were expanded with additional contextual attributes and have been analyzed along with "Bag-of-words", "Feature hashing" vectorization techniques and 4 machine learning models: "Support vector machine", "Convolutional neural network", "Logistic regression" and "Random forest". The outcomes of this objective are the results of 144 experiments represented as accuracy, precision, recall and F1 score metrics. "Logistic regression" model using "Feature hashing" vectorization technique on a "Fake news" dataset that is expanded with the main topic of author's message achieved the best accuracy score of 95,41%.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Petrauskas, Rokas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD210647789&prefLang=en_US