Complex asthma risk factor recognition from heterogeneous data streams

There are many studies regarding the relationships between environmental factors, particularly air pollution, and asthma exacerbation. Most of these studies ignore the potential confounding effects of a sequence of these factors with a specific time lag between them and asthma outbreaks. In this paper we present a new method for identifying consequential relations in the form of complex patterns between environmental factors and asthma attacks. Temporal structure and order relation between these data and their effect on asthma exacerbation comprise complex patterns called asthma risk factors. By extracting such patterns we create a risk prediction model that is important both for an asthmatic patient and public health. For experimental evaluations, we have collected pollution and meteorological data in Tokyo city and found 32 complex risk factor patterns that might result in asthma outbreaks. The experimental results show that extracted model has 71.15% precision.