Sleep efficiency is the measure of how effectively you utilize a person’s time in bed for sleeping, represented as a percentage of time asleep to time in bed. Maximizing sleep efficiency is important because it indicates better sleep quality, feeling more refreshed during the day, efficient time management, reduced sleep problems, and enhanced physical and mental health. This research aims to predict the value of sleep efficiency as a ratio of the time a person spends asleep to the total amount of time that person spends in bed.
The dataset we are using is obtained from the website Kaggle. It is part of a study conducted in Morocco by a group of Artificial Intelligence students from the National Higher School for Computer Science and Systems Analysis (ENSIAS), Morocco. The dataset has 452 rows, for 452 different test subjects, and each row records the statistics for one night of sleep. In this project, each data point also represents one night of sleep for one test subject.
Two different methods will be explored in this project: Polynomial Regression and Multi-layer Perceptron. Polynomial Regression is chosen due to its simplicity and efficiency with problems in which there are highly nonlinear relations between features and the label.
The Polynomial is performing better with a Mean-square error of 0.027.
The code for this project was written in Python 3.8 and requires the following libraries:
- Pandas
- Numpy
- Sklearn
- Matplotlib
- Seaborn
To run the code for this project, clone the repository and run the Jupyter Notebook file.