Please use this identifier to cite or link to this item:
|Title:||Sentiment analysis based on deep neural networks||Authors:||Zhang, Jingsheng||Keywords:||Engineering::Electrical and electronic engineering||Issue Date:||2021||Publisher:||Nanyang Technological University||Source:||Zhang, J. (2021). Sentiment analysis based on deep neural networks. Master's thesis, Nanyang Technological University, Singapore. https://hdl.handle.net/10356/152890||Abstract:||Now, with the rapid development of social media networks like Google, wikis, blogs, online forums, Twitter communities, Facebook communities, YouTube video platforms, and Tiktok short video platforms, the number and frequency of interpersonal read-write access interactions is increasing. Compared with existing methods, efficient machine learning models can perform accurate sentiment analysis on sample data. These models have many advantages, such as scalability, excellent real-time analysis capabilities, and consistent standards. Inspired by this, the main research content of this dissertation is sentiment analysis based on deep neural networks. This dissertation summarizes the sentiment analysis model based on deep neural networks. Develop deep neural network learning algorithms for sentiment analysis problems. The first part is Sentiment Analysis for Publicly Available Database. According to the existing database on the Internet, according to the results of various deep neural network models, compare the advantages and disadvantages of different classifiers, and observe and analyze the influence of different parameters on the model effect. According to different factors such as accuracy, variance, P value, F1 score and training time, the models are compared and improved. The second part is Sentiment Analysis on policy study. This section focuses on Singapore’s comments on the "tourism bubble" policy on social networking sites. For this specific policy, collect comments from people on the Internet and develop a deep neural network algorithm for sentiment analysis.||URI:||https://hdl.handle.net/10356/152890||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||EEE Theses|
Updated on Jun 27, 2022
Updated on Jun 27, 2022
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.