Development of operational methods to predict soil classes and properties in Canada using machine learning
Bibliographic record
Abstract
With a limited amount of arable land, worsening soil degradation, and plateauing of crop yields, the maintenance of limited soil resources to guarantee agriculturally productive soils and to support essential ecosystem services is critical in Canada and globally.The main objective of this thesis is to study and develop practical methods to cost-effectively renew soil and soil landscape information in Canada.Given the limited amount of pointbased soil data in Canada, machine learning-based predictive methods were studied for possible operational use.Machine learning methods require point soil data both for training purposes and validating the predictions.When soil data points are lacking, pseudo-point soil data are mined through soil survey polygon maps.In places where detailed soil surveys exist for soil class mapping, fully randomized pseudo-soil point data mining with random forest-based machine learning achieved a prediction accuracy as high as 74%.The prediction was further improved to a high of 78% by using an ensemble of multiple machine learners.Soil properties such as bulk density can be predicted either directly using point soil data or indirectly via predicted soil classes which are associated with reported soil property values.In this study, sampled soil bulk density values were used to predict soil bulk density across the study watershed.The predicted bulk density values come with uncertainty ranges, computed using residual kriging.Soil class prediction (and soil bulk density) may be carried out using environmental covariates from different sources.It is shown that where surficial geological material data are lacking, time series microwave remotely-sensed data, specifically Sentinel-1A synthetic aperture radar (SAR) imagery, can be used to delineate soil spatial patterns which are hypothesized to be linked to the spatial distribution of surficial geological materials.Through this study, a cost-effective iii work flow and solutions for predictive soil mapping needs in Canada were developed for operational use.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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 source (direct Gemma or distilled Codex), 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".