Ebola virus disease prediction using association rules

Abstract

Epidemic diseases pose a very big threat to Africa and the world at large, this has been as a result of abrupt or unforeseen disease outbreaks. However, a number of solutions have been proposed although most of them die as they are proposed. Mathematics, computing and data science have provided a number of techniques (including SIR models) to study the trends of epidemic diseases although it is still a challenge to predict the outbreak early enough to introduce control measures. With focus on data mining, association rules represent a promising technique to improve epidemic disease prediction. Unfortunately, when they are applied on a data set, they produce an extremely large number of rules. Most of such rules are irrelevant and the time required to find them can be impractical. A more important issue is that, in general, association rules are mined on the entire data set without validation on an independent sample. To solve these limitations, we introduce an enhanced Association rule algorithm (Enhanced Aprior) that uses search constraints to reduce the number of rules, searches for association rules on a training set, and finally validates them on an independent test set. The prediction significance of discovered rules is evaluated with support, confidence, and lift. Association rules are applied on a real data set containing medical records of patients with heart disease. In this study, associations rules relate Ebola symptoms and causes (Referred to as Attributes) with existence of the Ebola virus. Search constraints and test-set validation significantly reduce the number of association rules and produce a set of rules with high predictive accuracy. We exhibit important rules with high confidence, high lift, or both, that remain valid on the test-set on several runs. These rules represent valuable and interesting patterns.