Classification of Asthma Severity and Medication Using TensorFlow and Multilevel Databases

Escalating cost of treating chronic diseases demand that they be, to the extent possible, self-managed by the patients. In self-management of disease an imperative is to predict, the possible future state of morbidity (at time, T¹), given the present precursor conditions (at time, To) and expected precursor condition (at time, T¹). This paper reports the results of a study to evaluate the potential use of using TensorFlow and Inpatient Databases at national level and hospital level for predicting the asthma severity. Methods of Deep Neural Networks (DNN) have been deployed in classification of morbidity conditions, as well as treatment options. The results indicate that training a DNN to predict asthma severity level or the imminence of an asthma attack is possible.