Covid-19: Predicting affected patients using AI and vulnerability to environmental and demographic factors for India

EasyChair Preprint 7264

Abstract

This study uses climate and demographic variables such as location, humidity, minimum temperature, maximum temperature, population density were considered to predict Covid-19 cases across different states in India. In the present study a feed forward back-propagated neural model was used with Levemberg-Marquardt algorithm. The model showed good results with correlation coefficient (R) as 0.87 during training and 0.78 during testing between COVID-19 cases detected in actual and predicted by model (ANN). The sensitivity analysis of model revealed that humidity of the region was affecting the Covid-19 positive cases by 19.41% while day of year i.e. the day on which case was detected was contributing 18.39% while the demographic factor i.e. population density contributed 14.0%.

Keyphrases: ANN, COVID-19, Environment and demography, India, sensitivity analysis

@booklet{EasyChair:7264,
  author    = {Sunayana Chandra and Suresh Gurjar and Akash Priyadarshee and Vikas Kumar},
  title     = {Covid-19: Predicting affected patients using AI and vulnerability to environmental and demographic factors for India},
  howpublished = {EasyChair Preprint 7264},
  year      = {EasyChair, 2021}}