What's New in ACD/NMR Predictors
Version 12 to Version 2012
- Training Score tool implemented in the Spectrum and DB windows. The value of the Training Score is defined as the similarity of an atom's HOSE code compared to codes selected from corresponding structures in the Calculation Protocol. A higher Training Score indicates that an atom should be trained (prediction is not accurate).
- Utility Score tool added. The Utility Score (US) tool evaluates the usefulness of structures and their individual atoms for training other structures from the current database. It is expressed as a percentage ratio and calculated using the internal double-staged algorithm: first, the absolute usefulness of structures (considering all the atoms) is calculated; and second, competitive ranking is performed. The greater the US value, the more valuable the current structure is for training.
- Searches for NMR data on structure can be output directly to Microsoft Excel
- Improvements in the Calculation Protocol (CP); it now shows the number of coincided shells, and is resizable.
- Edit a shift value in the Spectrum window with a single-click on an atom (previously required a double-click)
- User DB can be updated from the Spectrum window without switching to the DB window.
| DB | Records | Shifts | Constants |
|---|
| C+H NMR | 336,821 | 4,204,295 | 811,315 |
| CNMR | 212,589 | 2,770,465 | 113,462 |
| HNMR | 228,794 | 1,956,762 | 739,163 |
| NNMR | 9870 | 23,320 | 4984 |
| FNMR | 19,438 | 38,947 | 36,217 |
| PNMR | 29,002 | 35,584 | 33,736 |
Version 11 to Version 12
- All NMR Predictors now include 1D NMR processing capability
- Internal database expansions
- HNMR Predictor: 210,700+ structures
- CNMR Predictor: 200,100+ structures
- NNMR Predictor: 9200+ structures
- FNMR Predictor: 17,000+ structures
- PNMR Predictor: 27,500+ structures
- Update or change structure in XNMR Predictor databases
- Neural network and incremental algorithms for 15N, 19F, and 31P NMR predictions
