Μεταπτυχιακές Εργασίες

Μόνιμο URI για αυτήν τη συλλογήhttps://pyxida.aueb.gr/handle/123456789/51

Περιήγηση

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω

Τώρα δείχνει 1 - 1 από 1

Hedge detection: an application on the wikipedia corpus
(2020) Agapiou, Marios; Αγαπίου, Μάριος; Athens University of Economics and Business, Department of Management Science and Technology; Chatziantoniou, Damianos; Spinellis, Diomidis; Louridas, Panagiotis
The purpose of this thesis is to develop a system that automatically detects hedges in Wikipedia articles, using weasel tags. The motivation behind this research project was to tackle the issue of ambiguity in Wikipedia articles, which could lead to the promo-tion of misleading information to the reader. This paper provides the general over-view of this task, including the extraction of the data, the classification methods that were used, as well as the evaluation metrics employed to examine the overall perfor-mance of these methods. In this thesis we experimented with machine and deep learn-ing models to apply the text classification. We implemented Support Vector Machine and XGBoost classifiers, and developed neural networks, such as Convolutional Neu-ral Networks (CNNs) and Recurrent Neural Networks (RNNs) with Long short-term memory (LSTM) architecture to complete this task. We then evaluated these systems against the best performing systems from previous studies that focus on this issue. Overall, we achieved notable results on our dataset, surpassing most hedge detection systems from previous studies, and thus proving the effectiveness of our methods.