Machine Learning Methods for Internet of Things in Medical Diagnosis

Internet of Things is increasingly getting popular in our daily life and also in many other fields. It can be caused by incredibly rising numbers of the devices connected to the Internet, like smartphones, tablets or laptops. Machine learning enables getting computers to act without being explicitly programmed. Furthermore, both of these fields are getting popular in the enormously essential field, which is medicine. The paper presents the analysis of chosen machine learning methods for cardiac arrhythmia diagnosis and choice of the best one with the highest accuracy as well as the construction of the device for ECG data gathering. Three machine learning methods were chosen for the analysis: Naive Bayes, K Nearest Neighbours and J48 decision tree. These methods were evaluated with the selected features. For this purpose, the MIT-BIH Arrhythmia Database was analysed and used features were elaborated.