Paper Title
Enhancing Sleep Disorder Diagnosis using XG Boost, Light GBM and Hybrid CNN-RNN Models

Abstract
Recognition of sleep disorders by means of sleep-stage classification and the development of a reliable diagnostic system have turned out to be highly essential in recent years in increasing both accuracy and speed in the diagnostic procedure. Patterns in conventional techniques, including polysomnography, are costly and time-intensive to determine and analyze; they are often subject to human error. High consistency with this process can be developed through the Machine Learning Algorithms. The background research is based on the Sleep Health Life-cycle Datasets that consists of demographic, lifestyle characteristics, and physiological variables. In the initial phase, it employs such benchmarks as Support Vector Machines, Random Forests, and k-Nearest Neighbors. It applies Advanced Deep Learning Techniques in the last phase. CNN's (Convolutional Neural Networks) extract valuable patterns in the shape or pickup of the information. Recurrent Neural Networks (RN-N's) have the ability to recognize how the data can evolve over time, particularly in physiological signals. The project as well employs genetic algorithms when it comes to training of the hyper-parameters to be used. In order to present an improved means of sleep stage Classification, it offers a contrast in the findings of a traditional ML tool in comparison with a DL tool. The aim of the study is to come up with a stable as well as a valid indicator that assists in improving the accuracy of health professionals on the issue of determining sleep disorders.