Applying Machine Learning to Design and Evaluate White Noise Recommendation System for Insomniacs

Abstract

Many sleep problems have occurred due to changes in the modern lifestyle. Insomnia is more serious than other sleep disorders. Insomnia increases the risk of depression, obesity, and cardiovascular diseases when it is not treated properly. Nowadays, most sleep therapies involve drugs that cause side effects on. For non-drug therapies, certain sounds assist sleep. However, as the sound is subjective, it is difficult to determine the sound suitable for each individual. This research designs an application that recommends white noise to insomniacs. For the application, we use machine learning technology for white noise recommendation and the design method for the user interface. For the experiment, we conduct a randomized controlled experiment and a five-day sleep experiment. This experiment verifies the effectiveness of the recommended white noise for sleep improvement. In addition to sleep assessment, we also use the system usability scale and semi-structured interviews to validate this system’s usability and willingness. The result shows that white noise improves deep sleep and reduces the time to fall asleep. Moreover, the usability score of this application is much higher than the passing score of the scale.