Selection and application of factors for forecasting the epidemic time and severity of Japanese encephalitis prevalence

For predicting median date or incidence rate in the prevalence of japanese encephalitis, the authors considered the 34 factors; namely, the climate, latitude, longitude, and date showing immunological positivity of hemoagglutination inhibiting reaction on 50% the number of swine, etc. To make the mean square residual the smallest which yields, in the case of calculating with the multiple regression equation, the most important and meritorious factors were selected from the factors mentioned above by the voluntary selection rule devised by us. Multiple regression equations were formulated for them. To predict the median date in the prevalence of japanese encephalitis in whole japan, we found that the three factors, i. e. the amount of rainfall in April, the average temperature in March and HI positive rate of swine till the end of july, were essential. And for the foreseeing of incidence rate in Okayama Prefecture, factors concerning mosquitoes were added; this resulted in the useful two factors, namely, common logarithm of the total number of Culex tritaeniorhynchus till july 20th and the rainfall in June.