USING ARTIFICIAL INTELLIGENCE WHAT CAUSES SENTIMENTAL ANALYSIS ON SOCIAL MEDIA

Abstract

With the use of sentiment analysis software, business leaders can keep tabs on public opinion as it
relates to certain brands, topics, and events. Millions of people utilize services like Twitter to share their
thoughts on a wide range of issues, and these technologies give dashboards to monitor the good, negative, and
neutral views expressed there. However, these techniques do not yet automatically extract explanations for
differences in sentiment, which makes it challenging for decision-makers to infer appropriate actions. In this
research, we begin by selecting the best performer among many Sentiment Analysis classifiers for brief texts by
comparing their performance. Then, we provide the Filtered-LDA framework, which much outperforms prior
techniques for deducing Twitter users' emotions. To identify potential explanations for shifts in opinion, the
system employs cascaded LDA Models with adjustable hyperparameters. Finally, a Topic Model with a high
Coherence Score is used to extract human-understandable Emerging themes by filtering out tweets on old
themes. At last, a unique Twitter dashboard for emotion reasoning is shown, which compiles the most
representative tweets for each potential explanation and displays them in a single place. The categorization
procedure would make use of machine learning and deep learning techniques. Sentiment may be tracked or
analyzed with the use of social media. This study provides a summary of the ways in which artificial
intelligence has been used to sentiment analysis on social media data to identify nervousness and hopelessness.