Machine Learning Classifiers for Asthma Disease Prediction: A Practical Illustration

Asthma is an inflammatory and airway-induced lung disease that causes breathlessness, wheezing and often life-threatening attacks. It is necessary for patients with asthma to predict severe exacerbation caused by uncontrolled asthma. So, there is a need for a framework that can predict the likelihood of asthma in a patient with maximum accuracy. In this paper, an effective mechanism for asthma disease prediction has been used by mining the data containing patients' previous health records. Four machine learning classification algorithms named as Naïve Bayes, J48, Random Forest and Random tree are used in these experiments to predict asthma disease at an early stage. The performance of all four algorithms is evaluated on various measures. Accuracy is measured over instances that are classified correctly and incorrectly. After experiments, results show that Naïve Bayes approaches 98% with the highest accuracy.