Sleep Health
What is a reason we might experience poor sleep quality?
Stress & anxiety, irregular sleep schedule, using electronic devices before bed, sleep disorders, poor sleep environment, poor diet, lack of physical activity, etc.
In this image, what type of correlation is shown?
Negative high correlation.
What was the dependant variable in the study?
Stress Levels.
Which regression model assumes a linear relationship between variables?
Linear Regression Model.
What is one topic this study investigated?
Sleep quality, etc.
What is a negative consequence associated with poor sleep quality?
Fatigue, mood disorder, cognitive decline, chronic disease.
Which sport is Tyler most likely to analyze using statistics?
Baseball.
What was an independent variable from the study?
Number of hours slept, body temperature, respiration rate, blood O2 levels, snoring and movement levels, etc.
What relationship was investigated using regression models?
The relationship between sleep-related variables and stress levels.
What were the most important variables for predicting stress levels?
Number of hours slept, blood oxygen levels, and respiration rate.
For adults, how many hours on average is optimal for best sleep quality?
7-9 hours per night.
What is one healthcare benefit of understanding sleep quality through data-driven analysis?
Stronger overall and mental health, early detection & prevention of sleep disorders, better diagnosis & prevention of sleep disorders, personalized sleep interventions, reduced healthcare costs and improved efficiency, enhanced integration of AI, wearable tech & sleep medicine, etc.
How many feature selection types were used in the study?
Three: wrapper methods, filter methods, and embedded methods.
What were two of the regression models tested in this study?
Linear, Ridge, Lasso, Random Forest.
Which variable was the least important for predicting stress levels?
Body temperature.
What is one sleep-related disorder?
Insomnia, Sleep Apnea, Restless Legs Syndrome (RLS), Narcolepsy, Parasomnias (i.e. Sleepwalking), Hypersomnia, etc.
What is one way we could further diversify or verify the dataset used in this study?
External validation, cross-population testing, time-based validation, integration of real-world data, etc.
How many feature selection techniques did this study employ?
Four: RFE, SelectKBest, Chi-Square Test, and Mutual Information.
Which regression model is similar to Linear, but prevents overfitting?
Ridge.
Which regression models were the most accurate in predicting stress levels?
Linear and Ridge.
What are some of the benefits of high-quality sleep?
Improved memory and cognitive function, better mental health and emotional well-being, increased physical performance and recovery, longer lifespan and reduced risk of chronic diseases, stronger immune system and heart health, improved social and work performance.
In multiple regression models, what is one way to reduce error?
Add more predictor variables.
How was each regression model evaluated for performance?
Root Mean Squared Error and R-squared tests.
Which regression model uses decision trees to capture non-linear relationships?
Random Forest Regressor.
Which regression model was the least accurate in predicting stress levels?
Lasso.