Correlation between Indoor Environmental Data and Biometric Parameters for the Impact Assessment of a Living Wall in a ZEB Lab.

Users' satisfaction in indoor spaces plays a key role in building design. In recent years, scientific research has focused more and more on the effects produced by the presence of greenery solutions in indoor environments. In this study, the Internet of Things (IoT) concept is used to define an effective solution to monitor indoor environmental parameters, along with the biometric data of users involved in an experimental campaign conducted in a Zero Energy Building laboratory where a living wall has been installed. The growing interest in the key theory of the IoT allows for the development of promising frameworks used to create datasets usually managed with Machine Learning (ML) approaches. Following this tendency, the dataset derived by the proposed infield research has been managed with different ML algorithms in order to identify the most suitable model and influential variables, among the environmental and biometric ones, that can be used to identify the plant configuration. The obtained results highlight how the eXtreme Gradient Boosting (XGBoost)-based model can obtain the best average accuracy score to predict the plant configuration considering both a selection of environmental parameters and biometric data as input values. Moreover, the XGBoost model has been used to identify the users with the highest accuracy considering a combination of picked biometric and environmental features. Finally, a new Green View Factor index has been introduced to characterize how greenery has an impact on the indoor space and it can be used to compare different studies where green elements have been used.