Abstract
With rapid increase in influenza A virus database, an important issue is whether the predictions are similar based on different datasets. Here we stratify 482 H3N2 hemagglutinins from influenza A virus in North America to different datasets. The predictions are made using logistic regression. The results show that the different datasets have significant impact on the predictions.
Keywords: Hemagglutinin, influenza, logistic regression, mutation, prediction