When the brain's blood flow is disrupted or decreased, a stroke happens. Because of a shortage of oxygen and nutrients in the brain, a stroke can result in the death of brain cells. By evaluating the effectiveness of predictive data mining approaches, a lot of effort has been done to predict different diseases. In this work, we assess various methods for predicting stroke in a dataset from a cardiovascular health (CHS) study. Hemorrhagic and ischemic strokes make up the majority of cases.
An ischemic embolism stroke typically occurs when a clot in bloodstream travels from the heart through the bloodstream to a smaller artery in the brain. Because hemorrhagic stroke happens when an artery in the brain leaks or ruptures, it is regarded as a different type of stroke. Stroke is one of the most serious diseases for persons over 65 and the second leading cause of mortality in the globe. It harms the brain in the same way that a "heart attack" harms the heart. A stroke can result in death as well as costly medical care and long-term impairment. A stroke claims the life of one person every four minutes, however up to 80% of strokes can be avoided with early detection or prediction.
Here, different machine learning classification approaches are used to create a classification model after dimensionality reduction using a principal component analysis algorithm and feature selection using a decision tree algorithm.
An ideal stroke predictions model with an accuracy of
95.27 was produced after studying and contrasting the effectiveness of categorization using various methods and the accuracy of the variant model.
Key words: Stroke, Cardiovascular health, Oxygen, Dataset, Model, Blood, Dataset.
|
