Assessment of machine learning and group contribution solvation parameter model descriptors for model retention in reversed-phase liquid chromatography and gas chromatography
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
• Modelling SoluteML and SoluteGC evaluated for RPLC and GC. • SoluteML estimated descriptors fit better than SoluteGC for chromatographic systems. • SoluteML and SoluteGC descriptors are not interchangeable with WSU descriptors. Abraham's solvation parameter model is a valuable tool for modelling reversed-phase liquid chromatography and gas chromatography systems. Except for the solute descriptor McGowan's characteristic volume, V, the remaining solute descriptors E, S, A, B, and L of the solvation parameter model are experimentally determined. Estimation approaches, machine learning, and group contribution methods are two alternatives to experimental approaches to estimating solute descriptors. In this work we evaluated the applicability of solvation parameter model solute descriptors estimated using machine learning and group contribution methods. Overall solute descriptors estimated using the machine learning approach fit better than solute descriptors estimated using the group contribution method for both reversed-phase liquid chromatography and gas chromatography systems studied in this work. For the studied methanol-water binary solvent system on a Luna C18(2) stationary phase model, coefficient of determination ranged from 0.982 to 0.953 when using machine learning estimated descriptors, whereas with group contribution estimated descriptors, models ranged between 0.923 and 0.943. For the studied gas chromatography models, coefficient of determination ranged from 0.995 to 0.987 when using machine learning estimated descriptors, whereas with group contribution estimated descriptors ranged between 0.941 and 0.977. However, both machine learning and group contribution descriptors did not fit in models as well as experimentally determined reference WSU descriptors.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it