Predictive Modeling of COVID-19 Patient Recovery Using Complete Blood Count Data

The COVID-19 pandemic originated in Wuhan, China, and has exhibited a rapid global spread. Efficient, trustworthy and readily available medical assessments of the disease's severity level can facilitate allocating and prioritizing resources to reduce death rates. The research aims to predict patients’ recovery status from the complete blood count dataset from Kaggle. The dataset was obtained from Dhaka Medical College Hospital, Bangladesh, from August 12 April to 31, 2020. The Hospital Ethical Committee authorized the collection of this data. In this paper, we have implemented recovery status prediction using the logistics regression and achieved a significant accuracy of 90.47%.