A novel framework for feature importance and sensitivity analysis in machine learning-based prediction of porosity, shrinkage, and bulk density in dehydrated products
Bibliographic record
Abstract
One key challenge in applying Machine Learning (ML) to drying process modeling is ensuring model robustness and reliability for downstream analysis. This study introduces a novel algorithm to assess the importance of quantitative and qualitative features in ML models developed to predict porosity, shrinkage, and bulk density during food drying. The approach emphasizes feature importance and model sensitivity to input combinations. Product type exerted the most substantial influence on model predictions, followed by temperature, type of drying method, and pressure. Pretreatment conditions and initial moisture content were critical for porosity and shrinkage modeling. Incorporating the additional inputs like initial bulk density, microwave power, or pretreatment increased model complexity without improving predictive accuracy. Sensitivity analysis demonstrated that a 10% increase or decrease in temperature altered predictions up to 22.2% and 27.3%, respectively. Pressure variations (±10%) led to accuracy changes of up to 14.3% for porosity and 19% for bulk density. Replacing apple and brown date data in the porosity model, carrot in the shrinkage model, and apple and potato in the bulk density model with other products caused accuracy changes of 26%, 63%, and 35%, respectively. Substituting data from experiments lacking pretreatment with those including pretreatment markedly altered model performance. Notably, freeze-drying and air-drying replacements induced the most pronounced accuracy shifts: 29.5% in porosity, 43.7% in shrinkage, and 109.2% in bulk density models. The findings highlight the direct impact of data distribution on model sensitivity, underscoring the importance of balanced datasets, normalization, and standardized measurement methods to improve ML model performance and generalizability.
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.
How this classification was reachedexpand
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 | 0.002 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| 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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".