n-Gram based language processing using Twitter dataset to identify COVID-19 patients