A genomic signal processing approach for identification and classification of coronavirus sequences

Corona disease has caused a variety of problems for people since it was formed and spread around the world. In this study, diagnosis and differentiation of this disease have been investigated in the form of genomic sequences. The proposed approach is based on a combination of several digital signal processing algorithms that include discrete Fourier transform and comb notch filter. More than 100,000 genomic sequences from different geographical locations and different variants have been tested in this research by various machine learning models. The use of KNN and SVM classifier models has resulted in the accurate diagnosis and differentiation of about 99% of coronavirus samples from the influenza virus. The proposed approach provides the possibility to generalize this method and improve machine learning models and get better results.