Pretrained Bootstrapped-Random Forest Tree (B-RFT) models for predicting gravimetric and volumetric hydrogen storage capacity in MOFs

This repository provides two pre‑trained machine‑learning models for rapid prediction of hydrogen storage capacities in metal–organic frameworks (MOFs). Both models were trained on 80% (78,956) of the curated global MOF dataset using bootstrap‑random forest techniques and are deployable in MATLAB.<b>Included Files:</b><code>UG_net_80_rand.mat</code>Predicts usable gravimetric capacity (wt. %)<code>UV_net_80_rand.mat</code>Predicts usable volumetric capacity (g H₂ L⁻¹)<b>Required Input Variables:</b>Single crystal density (g/cm³)Gravimetric surface area (m²/g)Volumetric surface area (m²/cm³)Void fractionPore volume (cm³/g)Large cavity diameter (Å)Pore limiting diameter (Å)Usage Example:% Load the pretrained modelUG= load('UG_net_80_rand.mat'); UV= load('UV_net_80_rand.mat');% Input variablesinp_val = [Single crystal density, Gravimetric surface area, Volumetric surface area, Void fraction, Pore volume, Large cavity diameter, Pore limiting diameter]% PredictionUG_pre=predict(UG.BaggedEnsemble,inp_val);UV_pre=predict(UV.BaggedEnsemble,inp_val);