The prediction models of anaphylactic disease

Abstract

Investigating the effect of common allergens on allergic disease is very important for human health. In this paper, we firstly propose the models for predicting the relationship between 39 common allergens and total IgE level. The total IgE level is utilized to evaluate the order of severity for allergic disease. In particular, we employ the linear fitting method and neural network based method to obtain the models with high prediction accuracy. The feasibility of the proposed models can be confirmed by testing two other independent data sets from hospital diagnosis record. Additionally, we obtain some useful medical conclusions.