Covid-19 Detection by Wavelet Entropy and Jaya

Wang, W.

The past year has seen a global pandemic of COVID-19, with extremely high transmission and mortality rates causing worldwide alarm and a significant negative impact on the global economy and human security. The asymptomatic COVID-19 patients have also made epidemic control very difficult. Using artificial intelligence technology to analyse CT images can locate infected patients quickly and precisely. Our research proposed a machine learning model based on wavelet entropy, single hidden layer feedforward neural network and Jaya algorithm, use K-fold cross-validation to report unbiased performance, which obtained a promising performance. The mean sensitivity was 73.31% ± 2.26%, specificity was 78.11% ± 1.92%, precision was 77.03% ± 1.35%, accuracy was 75.71% ± 1.04% and F1 score was 75.10% ± 1.23%. Matthews correlation coefficient of 51.51 ± 2.07%, and feature mutual information of 75.14% ± 1.22%. Our research proved the importance of AI technology for the medical field to a certain extent. © 2021, Springer Nature Switzerland AG.