Through study of research papers related to studies conducted around Sleep Apnea syndrome, project identified by viewing the data of the population affected by this particular disease and several studies that it find common factors including Age,Gender, Diabetic Condition, Drug Symptoms,Breathing Disorder, Location and Spo2 levels that contribute towards Sleep Apnea disorder.
Hence find data related to the disease and prepared it neat and clean and then taking input of User data including manual entrance of age,gender,diabetic status, Drug Addiction and Breathing Disorder, Location through User’s Gps and collection of spo2 through an IOT-based Oximeter sensor into the Sleep Apnea Android App, prepared a Machine Learning Model and trained it using the collected data to predict whether a patient is suffering with Sleep Apnea Disorder or not and displayed the results to the user on Android Application.
1. To build an IOT based sensor to be used as an oximeter that inputs user Spo2 levels and BPM level and saves it to a database.
->To input Spo2 levels developed an IOT-based Oximeter Sensor using MAX30100 Pulse Oximeter Sensor and ESP8266 WiFi MCU(NodeMCU) setup circuited on breadboard and connected to NodeMcu Arduino Ide via USBports. These readings were saved as data in our Sql database model on Xampp Local server.
->To find GPS location,the project identifies the device's physical location using Android Location Service. In the project, User permission is given and a function is created to enable GPS service and retrieve location information in the form of Location_Manager variable through which identify and retrieve longitude and latitude details. Using these coordinates, the project finds the State and City of the device and displays them on User Interface.
3. To build and train a machine learning model to predict whether a person suffering from sleep apnea or not using suitable cleaned data.
->To predict whether a person is diagnosed with Sleep Apnea found machine learning to be an effective tool to train the data and hence developed an machine learning model.For this collected information on various models that can be used for binary classification of a dataset among them decided to work on Neural Networks
4. To build an Android Application that takes user data from a database and print the result after successful prediction by Machine learning model.
->The Android Application has been developed using Java, XML and SQLITE database which is locally stored in the Android.The App comes with Register,Login and Logout Activities with User Session saved using SharedPreferences. Navigation Bar has three Fragments i.e. Home Fragment, Predict Fragment and Profile Fragment.
