Ilya Khodov LabNew paper published in JCIM (@ACSPublications)!
We investigated whether accurate prediction of transition-metal NMR chemical shifts really requires computationally expensive 3D structures and quantum-chemical calculations.
Using nearly 2,000 experimental measurements for Mn, Fe, Nb, and Mo complexes, we found that machine learning models based only on 2D molecular descriptors can achieve surprisingly strong predictive performance. https://doi.org/10.1021/acs.jcim.6c01787
We investigated whether accurate prediction of transition-metal NMR chemical shifts really requires computationally expensive 3D structures and quantum-chemical calculations.
Using nearly 2,000 experimental measurements for Mn, Fe, Nb, and Mo complexes, we found that machine learning models based only on 2D molecular descriptors can achieve surprisingly strong predictive performance. https://doi.org/10.1021/acs.jcim.6c01787
