Sentiment analysis of microblog text based on sentiment dictionary

Abstract

This study introduces a method for sentiment analysis applied to microblog texts, leveraging a sentiment dictionary. Focused on discerning sentiments as positive, negative, or neutral, the research incorporates a sentiment dictionary tailored for microblog content. The experiment employs a dataset pertaining to the new coronavirus epidemic collected from a microblogging platform for testing and compares the outcomes with those of existing methods. Key enhancements include the integration of a microblog-specific emotion dictionary and the formulation of semantic rules designed for the nuances of Chinese text. Results demonstrate improved accuracy in text emotion recognition, outperforming traditional methods and achieving notable accuracies for positive, negative, and neutral sentiments. Objective: To develop and implement a robust sentiment analysis model for microblog text, leveraging a comprehensive sentiment dictionary to accurately classify the emotional tone and polarity of textual data. The model aims to enhance the understanding of user sentiments on microblogging platforms by examining the nuances of text and identifying positive, negative, or neutral sentiments. The analysis will facilitate the extraction of insights from large volumes of user-generated content, enabling applications such as social media monitoring, brand perception analysis, and opinion tracking.