Application of XGBoost Model for Predicting Isobaric Heat Capacity of Liquid HC, HFC, and HFO Refrigerants

This article presents a machine learning model for accurately calculating the isobaric heat capacity of pure HC, HFC, and HFO refrigerants in the liquid phase. Initially, a comprehensive evaluation was conducted on 1687 experimental data points encompassing various refrigerants. Subsequently, distinct linear algorithms, nonlinear algorithms, and ensemble algorithms were individually trained and tested. After careful comparison and analysis, XGBoost emerged as the most effective algorithm for this task. The utilization of XGBoost in predicting and computing the isobaric heat capacity of refrigerants demonstrated excellent agreement with experimental values. The resulting <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{R}^{2}$</tex> value was found to be 0.9734 while RMSE stood at 0.1267 <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{kJ}/(\text{kg}\cdot \mathrm{K})$</tex>.