Machine Learning-Based Emerging Technologies in the Post Pandemic Scenario

ABSTRACT

The entire world is now fighting to overcome the after-effects of the COVID-19 pandemic. The most challenging factor is handling the post-pandemic scenario in various sectors such as medical, educational, civil services, business, etc. Every sector put its efforts in the best possible way to cope up with these unprecedented times. The majority of the sectors came up with the idea of work from home, robotics, online payment system, entertainment, meetings, webinars, etc., to continue with the smooth working environment. To manage this crisis, it is essential to be equipped with appropriate tools and technologies. These technologies may differ from sector to sector to satisfy their essential requirements. Among these, the educational sector is the only domain in which the entire transition happened from the classroom model to a fully virtual model of service. The digitalization of the educational system ensured the continuous delivery of education without any requirement of physical presence. The proposed idea uses data science techniques to understand the crisis in a better way and improvise the quality of education. In the medical sector, there are various techniques like online consultation with doctors, telemedicine and chatbots. Augmented Reality solutions provide a better platform for convenient, smooth and safe doctor-patient interaction. Data science techniques are implied to build an efficient novel approach for early diagnosis in medical sectors and reduce the mortality rate. Similarly, in other sectors such as business and civil services, there are betterments going to be brought about in a virtual environment. The segment which is facing the repercussions of this deadly virus is the common man, especially farmers, small businessmen, daily wagers, elders, children, and patients. Considering the current situation of the common public, one should try to make a brave new world, which is more adaptable, sustainable and stronger to mitigate the risks. This system includes a CNN model that classifies if a pair of lungs is normal or has pneumonia. The CNN model provides 87.71% accuracy, 12.29% miscalculation rate, 80.45% precision, 80.94% sensitivity, 91.20% specificity, 80.03% F-1 value against the testing dataset having a loss value of 0.5096. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.