Page 69 - Read Online
P. 69
Page 14 of 21 Wen et al. J. Mater. Inf. 2025, 5, 30 https://dx.doi.org/10.20517/jmi.2024.102
Figure 6. The true values and ML predicted values for (A) Hole reorganization energy, (B) Solvation free energy, (C) Maximum
absorption, and (D) LogP based on the GBDT model, respectively. ML: Machine learning; GBDT: gradient boosted decision tree.
parallelized tree learning, significantly accelerating computational speed while maintaining precision. In
contrast, conventional GBDT lacks such systematic optimizations, resulting in suboptimal efficiency-
accuracy trade-offs. These innovations collectively enable XGBoost to achieve a balanced and robust
performance in both accuracy and scalability.
Validation of ML predictive models
To further assess the performance of the model on a new dataset, seven reported HTMs, including spiro-
OMeTAD, DTPC8-ThDTPA, DTPC13-ThTPA, DTP-C6Th, TPA-TVT-TPA, YZ18, and YZ22, were
selected as test samples. The RF, GBDT, and XGBoost ML models were then employed to predict the hole
reorganization energy, solvation free energy, maximum light absorption peak, and hydrophobicity of these
molecules. The generalization ability of the models was evaluated by comparing the predicted values from
