Volume 03: Proceedings of 11th International Conference on Applied Energy, Part 2, Sweden, 2019

Relative Humidity Estimation: Machine Learning Approach–Random Forest-Based Prediction Model Kinza Qadeer, Ashfaq Ahmad, Muhammad Abdul Qyyum, Moonyong Lee1


Relative humidity (ɸ) is considered a major parameter during the designing of HVAC (Heating, ventilation, and air conditioning) systems. Generally, HVAC engineers use a psychrometric chart to observe and estimate the air quality parameters. Nevertheless, high skills are required to make rigorous and accurate reading from the psychrometric chart and the “human error” is an added factor that can lead to big disasters. Therefore, rigorous and user-friendly estimation of air quality parameters is still an ongoing issue. In this context, we are going to implement the state-of-the-art “Machine learning” technique to develop a simple, robust, and rigorous predictive tool for the estimation of relative humidity. A well-proven approach i.e., the random forest (RF) is employed to train the model for robust estimation. It was found that the mean absolute deviation was 54.3% lower than that of well-known ordinary least square (OLS) regression method.

Keywords Relative humidity, random forest, prediction, ordinary least square, HVAC engineers

Copyright ©
Energy Proceedings