Detection of Panic Attacks Using Machine Learning Models

Abstract

Panic attack is psychological condition, is primarily caused by experiencing sudden mental or physical symptoms such as sweating, sudden jump in heart rate. However, the few similar symptoms also happen during Heart attacks. So, by default most of the patients are categorized as Heart patients even they have a panic attack which cause confusion. This not only create nancial losses but also threat to human life and safety. This paper presents a panic attack prediction approach to counter such cases by initial symptoms it can detect the person is having a panic attack or not. To counter this situation this study employes machine learning models such as SVM, Nave-Bayes and Random Forest to predict the panic attack from mixed readings of panic attack and heart attack data collected by wearable devices and classify if the person is having a panic attack. Utilizing the modern techniques of panic attack prediction such as these model's accuracy is improved with a result of 94% and enhanced the time e ciency. Furthermore, this framework enhances the process of diagnosing a panic attack and can save e cient resources in initial stage of diagnostic for heart attack.