Machine Learning Prediction of ¹H NMR Chemical Shifts in Zinc Complexes

Title: Machine learning prediction of ¹H NMR chemical shifts in zinc complexes — Dataset Description:This dataset contains all data and notebooks used in the study “Machine learning prediction of ¹H NMR chemical shifts in zinc complexes.” It includes: Raw DFT calculations for 45 zinc complexes, including geometry optimizations and solvent-specific NMR shielding calculations. Experimental ¹H NMR data, including raw spectra, processed spectra, and peak assignments for validation. Compiled XYZ files used for ML descriptor extraction (SOAP) and clustering. The notebooks reproduce the full machine learning workflow from raw data: H_Clustering.ipynb — performs clustering of zinc complexes using the merged XYZ structures. HNMR_lgbm.ipynb — extracts SOAP descriptors from hydrogen atoms, trains LightGBM models to predict ¹H NMR chemical shifts, and validates predictions against experimental data. Dependencies / Required Python modules:python, numpy, pandas, scikit-learn, lightgbm, dscribe, ase, matplotlib