| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 9.21 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
O YouTube é uma plataforma de partilha de vídeos que permite obter apreciações por parte dos utilizadores. Os utilizadores podem demonstrar uma opinião positiva ou negativa a partir de determinados botões, ou podem manifestar essa opinião a partir da escrita de um comentário sobre o vídeo. O YouTube permite saber quantas apreciações positivas e negativas um vídeo teve, mas não permite saber de forma automática se os comentários são globalmente positivos ou não. A análise sentimental de textos permite determinar, computacionalmente, sem comentário é positivo, negativo ou neutro.
No trabalho descrito neste documento foi desenvolvida uma aplicação para recolha de dados disponíveis na plataforma YouTube. Efetuou-sea análise sentimental de comentários extraídos do YouTube com recurso a técnicas de processamento de linguagem natural. Também foram analisadas diferenças na escrita dos comentários. Consideraram-se dois grandes grupos de tipos de vídeos: vídeos técnicos e não-técnicos, os primeiros geralmente de um cariz educativo e os segundos mais vocacionados para o entretenimento.
Os dados foram analisados e as semelhanças e diferenças encontradas entre diferentes tipos de vídeos estão reportadas. Verificou-se que os comentários de vídeos não técnicos, em geral, apresentam a sua positividade e negatividade de forma mais acentuada.
YouTube is a web platform for sharing videos which allows users to like or dislike them. Users can demonstrate a positive or negative opinion by clicking the respective buttons, or can express their opinion by writing a comment on the video. YouTube lets us know how many positive and negative appreciation a video has, but does not allow to automatically know if the comments are positive or not. The sentimental analysis of texts is the way to know, computationally, if a comment is positive, negative or neutral. In this work, an application was developed to get YouTube available data. The goal of this study was to perform the sentimental analysis of the extracted video comments using natural language processing techniques. It was also intended to analyze the behavior of the users and their type of writing in different types of videos. It was considered two groups of types of videos: technical videos and nontechnical. Technical videos have an educational nature and the nontechnical are geared for entertainment. The collected data were analyzed and the similarities and differences between different types of videos are reported. It was found that the comments of nontechnical videos usually have their positivity and negativity more accentuated.
YouTube is a web platform for sharing videos which allows users to like or dislike them. Users can demonstrate a positive or negative opinion by clicking the respective buttons, or can express their opinion by writing a comment on the video. YouTube lets us know how many positive and negative appreciation a video has, but does not allow to automatically know if the comments are positive or not. The sentimental analysis of texts is the way to know, computationally, if a comment is positive, negative or neutral. In this work, an application was developed to get YouTube available data. The goal of this study was to perform the sentimental analysis of the extracted video comments using natural language processing techniques. It was also intended to analyze the behavior of the users and their type of writing in different types of videos. It was considered two groups of types of videos: technical videos and nontechnical. Technical videos have an educational nature and the nontechnical are geared for entertainment. The collected data were analyzed and the similarities and differences between different types of videos are reported. It was found that the comments of nontechnical videos usually have their positivity and negativity more accentuated.
Description
Keywords
Análise sentimental Processamento de linguagem natural YouTube Comentários Polaridade Sentiment analysis Natural language processing Comments Polarity