The asymptomatic ratio (the proportion of asymptomatic infected cases among the total infected cases) is an important indicator in public health. However, symptoms manifested by a case infected with a pathogen may result from infections other than the specific pathogen of interest. When a case is infected with multiple pathogens, it can be difficult to distinguish which pathogen is causing the symptoms. These conditions complicate the estimation of asymptomatic ratios. We used influenza serologic data from Taiwan to test a novel log-linear binomial regression model that estimates pathogen-specific asymptomatic ratios for influenza subtypes. We find that 75% of the H1N1 subtype and 65% of the H3N2 subtype were asymptomatic. Asymptomatic ratios help to quantify the magnitude of asymptomatic persons who may be capable of spreading pathogens to others.