Supervised Multimodal Emotion Analysis of Violence on Doctors Tweets
Parth Vyas, Falak Sharma, Akhtar Rasool, Aditya Dubey · 2021 Fifth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC) · 2021
With the advent of novel coronavirus pandemic doctors, health workers, and the government too, are trying their best of their capacity to deal with contemporary situations. It is genuine that when a person’s close one is lost, they will react vociferously but accusing the doctors and workers and harming them is also morally indignant as the person saving so many lives his/her own life is in danger. With the boom of technology and how the world has come so close on social media, many social media users are expressing their views in either the support or opposition of the saviors of this pandemic, the doctors and the health care workers. These views of people are enough to create a good or bad impression of any doctor in minds of people and can even create a hostile behavior for that doctor by others, analyzing the stand of the person towards the ongoing violent situation towards workers using a multimodal emotional analysis combining both visual and textual data. This paper uses a Multimodal Transformer model which combines both visual and textual data is the sole purpose of this paper. Apart from the main aim, the paper will also explain whether in social media more information has been carried out by a text or more information can spread through images posted on social media. The paper will explain the use of appropriate loss function for imbalanced data also.