A Machine Learning Enabled Mobile Application to Analyse Ambient-Body Correlations

Ambient factors and living conditions have the capacity to cause mental and/or physical diseases if they are not properly managed. This paper studies the impact of ambient factors on body parameters. For this, we design a mobile application that collects ambient features and body data samples via a Bluetooth-enabled sensory system to train machine learning prediction models. It uses random forest, linear regression, and boosted tree techniques to predict the body parameters based on the ambient conditions. The machine learning models are evaluated and compared in terms of prediction accuracy and RMSE to find the best-fitted prediction approach. According to the results, the boosted tree model outperforms random forest and linear regression and gives the best prediction accuracy.