Analysis Of Multimodal Data On Social Media Using Deep Learning Techniques

Abstract

Contextual text mining known as sentiment analysis identifies and extracts subjective information from the source content. It aids in the detection of sentiments that are good, negative, neutral, etc. It helps companies monitor internet debates in order to learn how the public feels about their brands, goods, and services. However, the only metrics generally utilized in social media stream analysis are straightforward sentiment analysis and count-based metrics. This is analogous to simply scratching the surface and leaving out those priceless discoveries that are just waiting to be made. Sentiment analysis is quickly evolving into a crucial tool to track and comprehend the sentiment in all types of data because people express their thoughts and feelings more freely than ever before. This project's sole objective is to use various latest AI techniques to categorize various sentiments present in audio and text forms into categories like humorous, offensive, and sarcastic. Using datasets with audio files and image files, we trained the model, then we tested it using the test data.