<![CDATA[This research compares the performance of two sentiment analysis algorithms, namely VADER (Valence Aware Dictionary and Entiment Reasoner) and RoBERTa (Robustly Optimized BERT Pretraining Approach), using a dataset of public opinions regarding climate change on twitter. Analysis is carried out to determine the sentiment distribution of the tweets described, whether they are positive, negative or neutral. In addition, this research identifies the keywords that appear most frequently from the collection of tweets that have been analyzed. Time series analysis was also carried out to see the distribution of sentiment over 12 months. The relationship between the two models was evaluated using matrix scatter plot analysis for tweets per two months, to assess the correlation and consistency of sentiment results between VADER and RoBERTa. The results show that VADER is more effective in situations that require rapid responses to changes in public sentiment, while RoBERTa is superior in in-depth analysis of more complex and ambiguous content.]]>
Copyrights © 2024
