Using an ensemble learning approach in digital soil mapping of soil pH for the Thompson-Okanagan region of British Columbia
Bibliographic record
Abstract
Information on the spatial distribution of soil pH is essential for assessing soil quality and soil productivity. Digital soil mapping (DSM) is commonly used to predict soil characteristics over various types of landscapes. Over the past decade, researchers have made progress using machine learning techniques to provide reliable predictions of soil properties with limited data. DSM studies often use a single learning approach, which is constructed with a machine learner that systematically extracts soil–environment relationships from a large database, whereby a fitted model is used to predict soil information in an unmapped area. The practice of using an ensemble learning approach, especially one that combines several base learners, has rarely been tested in DSM. We developed a workflow for using an ensemble learning algorithm to predict soil properties for the Thompson-Okanagan region of British Columbia, Canada. Here, we focused on soil pH and tested a variety of base learners. Base learners with high prediction accuracies were then used to construct a SuperLearner (SL) to extract the complex relationships between soil properties and environmental variables. The fitted SL was then used to predict soil properties at 25 m spatial resolution at three depth intervals (0–5, 5–15, and 15–30 cm). Prediction accuracies were assessed using an independent test dataset, which indicated that the SL had a similar prediction accuracy to the best individual base learners. Using the heterogeneous ensemble learning approach with a weighted average stacked generalization process eliminated the need to choose the best base learner.
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 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 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".