Improving DFT Calculation Accuracy for Y-NO Bond Energies
Author Information
Author(s): Li Hong Zhi, Tao Wei, Gao Ting, Li Hui, Lu Ying Hua, Su Zhong Min
Primary Institution: Northeast Normal University
Hypothesis
Can a generalized regression neural network improve the accuracy of DFT calculations for homolysis bond dissociation energies of Y-NO bonds?
Conclusion
The GP-GRNN approach significantly improves the accuracy of calculating homolysis bond dissociation energies, reducing the RMS error from 5.31 to 0.31 kcal mol−1.
Supporting Evidence
- The GP-GRNN approach reduced the RMS of calculated homolysis BDE from 5.31 to 0.31 kcal mol−1.
- The study involved 92 organic molecules containing Y-NO bonds.
- Statistical methods like GRA and PCA were used to optimize molecular descriptors.
Takeaway
This study shows a new way to make calculations about how strong certain chemical bonds are, which helps scientists understand how these bonds work better.
Methodology
The study used a generalized regression neural network (GRNN) combined with grey relational analysis (GRA) and principal component analysis (PCA) to improve DFT calculations for 92 organic molecules.
Limitations
The prediction accuracy may be limited by the small training set size and the features extracted during the training process.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website