Predicting Sleep Disorders for Improved Health Care—A Comprehensive Study

All around the world there are wide range of people who are suffering from the sleeping disorder problems such as insomnia and apnea. The people who are suffering from this kind of sleep disorders are highly impacted with both physical and mental illnesses. Sometimes the people don’t know by what they are suffering from, in that case the people should consider a doctor for diagnosing the problem. Our idea is that can a person identify whether he is having disorder or not on his own? Well, it is possible by the combination of both AI technology and health care. In our study, we used supervised machine learning algorithms and built a model that which can guide the user whether he is having sleep disorder or not. The model is trained with the dataset which we collected from Kaggle and we started our study by analyzing the data and we found crucial insights which are very helpful for building our model. After verifying all these insights and with satisfactory results, we trained our data with various supervised classification algorithms, among all of them the both Naive Bayes and random forest had given better accurate and promising results and we also applied XGBoost for improving the performance of our model. Additionally, we further tested our model with real-world data which we collected from our classmates, faculty, and other individuals irrespective of their age groups through Google Forms. Finally, we developed a live application with the random forest model because, as though the Naïve bayes model performs well, naïve bayes model will give the outcomes basing on the probabilities. Based on the probabilities it is not preferable to guide a person regarding his health state. So, instead naïve bayes we have preferred random forest model for our application and deployed it on the Streamlit platform. Our main aim is to reduce the traffic to the doctor and to make the person well aware of his health state. As if a person is not suffering from the disorder, then it’s waste of time to the doctor, instead in that time he might can be treated to other person who is really suffering from the disorder.