Predicting Stroke by Combination of Sequence Pattern Mining and Associative Classification

Abstract

Stroke is a medical emergency that needs immediate medical attention. It is the third cause of death in the world and is the first cause of death of elderly women in Thailand. Stroke needs to be predicted in order to prevent people from the disease and to prepare proper medical treatments for the patients. A number of research works tried to study factors, such as blood pressure, smoking, and cholesterol, for predicting stroke. Unlike the previous works, the association of disease sequence is combined with factors for predicting stroke in this paper. The association is represented in the form of class sequential rules which demonstrate the association of diseases and factors leading to stroke. The combination of sequential pattern mining and associative classification is proposed as a method for generating class sequential rules. The experimental results show that the proposed technique gives high performance for the prediction. In addition, this paper shows top ten association of the disease and factors leading to stroke.