A Novel Approach to Predict the Real Time Sentimental Analysis by Naive Bayes RNN Algorithm during the COVID Pandemic in UAE

Abstract

At present, each and every part of the globe are facing a COVID crisis, which is affecting an individual physically, mentally and on the other hand, it is affecting the nation economically. Also, the unemployment scenario will be at its peak in the upcoming years as reported by UNGA. To combat this scenario, all the country are working on fostering their fiber network and so the sectors apart from the manufacturing, will tend to work from their home and contribute to the economy. But there are many problems arising to implement this culture practically, since it affects the mindsets of the people who have to endure this transformation within a very short span. Hence, in this research work, it has been decided to focus on this current issue for which the usage of certain apps in UAE such as zoom, totok, botim for internet calling have been identified since this is the only way of connectivity with the outside world. To perform this analysis, the tweets from December to July have been collected by converting the image to text and analyzed using two algorithms such as Naive Bayes Classifier (NBC) and Recurrent Neural Networks (RNN). The sentimental analysis found that 630 tweets were positive and people in UAE feels secured, satisfied and internet calling is very useful for them in the prospect of work, education, etc. Only 48 tweets has negative impact because people feel little bit harder in sudden change of culture with in short period of time and 155 tweets has impact that both positive and negative were view and said to be natural. The study found that NB (84%) is more accurate, user friendly and takes less time than RNN (79%) to perform the analysis. Finally, the sentimental analysis reveals that people in UAE were accepting the new culture of internet calling and it is useful for them in the prospect of work and education.