Project Title: Analyzing Sleep Data With Machine Learning
BASIS Advisor: Mr. Hansen
Internship Location: Somnoware Healthcare Systems
Onsite Mentor: Mr. Mathew Mammen
My internship will experiment with machine learning and healthcare. I will be working for a company named Somnoware that enables chronic obstructive pulmonary disease and sleep apnea care providers to reduce cost and improve patient outcomes. I will be studying machine learning and taking sleep data to do a predictive analysis and predict what diseases a person is prone to based on that data. I aim to find which machine learning method is most effective in accurately predicting obstructive sleep apnea and will learn about probabilistic graphical models, pattern recognition, and numerical computing. Then, I will gain familiarity with Octave and MATlab, collect data from classmates using Fitbits, and train the test data. I will decide which modeling is most accurate, test the model, and debug any efficiency problems. I expect that neural networks will be the most efficient because they can have many layers with non-linearities and are effective at modelling highly complex non-linear relationships. I hope to discover how machine learning can transform the way that we sleep and the accuracy of analyses made by doctors in regards to sleeping disorders. The sleep data that I collect from my classmates can be useful in showing other teenagers how you can plan sleep in order to minimize disease risk.
My Posts
Week 11: Finalizing Details
Hello everyone, I’ve been practicing how I’m going to present the powerpoint (eye contact, memorizing most of the material, projecting my voice, etc.). I’ve decided that I’ll make a poster as my final product and include various graphs and charts that can organize my data and make it easily understandable. My project has been completed […]
Week 10: Presentation
Hello everyone, I’m currently preparing for the presentation by putting together slides with graphs that represent the work I’ve been doing. In addition, I’ve made sure to include context on why my project is important and research that exists regarding OSA. As I’ve stated in a previous blog, I want to include less jargon and […]
Week 9: Preparing for the Presentation
Hello everyone, I’m coming towards the end of my ML course and am brainstorming ideas about how to present my project. I have a call scheduled with my external advisor later this week in which we will finalize most of the project. I’ll let you guys know how that goes. As usual, I’ve been reading […]
Week 8: Assignment
Hello everyone, Unfortunately, I fell sick for the first half of the week and that wasn’t exactly pleasant. However, I did watch MATLAB tutorials on youtube and read some articles about similar projects. In my course, I recently learned about Gaussian Distribution, how to build an Anomaly Detection System, and even how to predict movie […]
Week 7: Direction
Hello everyone, Since presentations are just around the corner, I’ve been brainstorming how to mold my project in a way that it could be easily presented. Yes, the code may be interesting for some but I want to make my presentation contain less jargon and more facts that can be interesting for all. As I […]
Week 6: New Datasets
Hello everyone, In addition to the sleep dataset that I have received from the startup, I was looking online for more datasets and came across the National Sleep Research Resource (NSRR). I spent time looking at various studies with hundreds, if not thousands, of subjects and different age groups. One that caught my eye was […]
Week 5: Articles and Coursework
Hello everyone, This week wasn’t as interesting as last week but I still got some good progress done. I needed my external advisor to answer some questions in regards to the data but he was quite busy. This meant that there was slow progress on the project itself but I used the time to progress […]
Week 4: Existing Sleep Questionnaires
Hello everyone, Since I’ve decided to focus on accessibility, I’ve been looking into existing sleep questionnaires that can be done online in order to know if one should go to the doctor for a sleep test. The four main sleep questionnaires that I studied are Berlin, Epworth Sleepiness Scale, STOP, and STOP-Bang. They all focus […]
Week 3: Accessibility
Hello everyone, To diagnose OSA without a professionally conducted sleep test, I have to focus on non-polysomnographic variables. This means that the tool will not be as accurate as I’d like, but will still achieve the purpose of telling someone whether there is a decent chance that they have OSA and if they should go through […]
Week 2: Learning about the Barriers
Hello everyone, This week I looked into research papers done on Obstructive Sleep Apnea and read about experiments, both successful and unsuccessful, that implemented machine learning to diagnose OSA. Previously, I envisioned creating a project that could study existing patterns in the sleep statistics of those already diagnosed and make a diagnosis on a new […]
Week 1: Delving into Machine Learning
Hello, my peers and other inquisitive persons, I’ve spent the past week laying a solid foundation for my upcoming pursuits. From Dr. Andrew Ng’s Machine Learning course on Coursera to various research papers online, I’ve been studying a variety of important topics that will help me reach my final goal of predicting Obstructive Sleep Apnea […]