Stroke Factor Analysis based on Granger Causality Test

Abstract

Stroke is a common brain disease that happens when brain vessels get blocked. As one of top-rated reason for human death, stroke not only leads to death but also causes huge economic damage and occupies a lot of medical resources. Every year 610,000 people died in US because of stroke. The economy loss is up to 30 billion dollars. This paper tries to predict and find the relationship between stroke and other related variables. The dataset from Kaggle collects 5110 participants' data about stroke with other health or habitats’ relative variables like gender, age, bmi, average glucose level, whether have hypertension, smoking status, to a total of 12 variables. With the use of the VAR model and Granger Causality test, 10 different pairs of VAR models are created and compared. The result of the final test shows that 9 variables have Granger Causality with stroke while only 1 attribute is not.