Challenges and Next Steps for Land-use Regression Models
Bibliographic record
Abstract
S-30B1-5 Background/Aims: This talk addresses lessons learned, challenges, and next steps in the application of the land-use regression (LUR) model approach for air pollution exposure assessment. Specifically, we will discuss creating national LUR models, developing LUR models for developing countries, and matching the spatial resolution of LUR models with microenvironment location information. Methods: We developed national LUR models using fixed site monitors for the United States and Canada (obtained from the US Environmental Protection Agency and NAPS respectively), land use characteristics, and satellite pollutant measurements. Results: For the United States, models were developed for annual average (R2: 0.78) and seasonal (R2: 0.73–0.77) concentrations of NO2. For Canada, models were developed for annual average concentration of particulate matter 2.5 (PM2.5) (R2 = 0.44) and several gases: NO2, Benzene, Ethyl benzene, and Butadiene (R2: 0.62–0.69). Within-city variability is predicted reasonably well for the US model, but requires further work for the Canadian model. A second area of current work involves LUR in Delhi, India, for PM2.5, black carbon, and particle number concentrations. Preliminary results suggest differing spatial patterns in India compared to typical North American contexts and the need for different predictor variables—for example, in high-density low-income neighborhoods where in-home combustion sources are predominantly solid fuels, PM2.5 concentrations may be higher away from a roadway than near a roadway. Our third area of current work is to improve microenvironment location information for exposure estimation. This work builds on prior mobility-based exposure models. CAREX Canada has developed “daytime” location estimates from satellite data at block and dissemination areas, to calculate population exposure estimates using LUR models. Following on earlier work using geocoded activity diaries and logging GPS, we are working with smart-phone GPS technology to estimate individual exposures based on location. Conclusion: The use of national scale LUR models shows promise. Expanding to developing counties and improving the microenvironment information present a set of challenges to be addressed.
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 | 0.031 | 0.063 |
| Meta-epidemiology (narrow) | 0.003 | 0.002 |
| Meta-epidemiology (broad) | 0.003 | 0.003 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.001 | 0.003 |
| Scholarly communication | 0.006 | 0.011 |
| Open science | 0.006 | 0.004 |
| Research integrity | 0.005 | 0.012 |
| Insufficient payload (model declined to judge) | 0.012 | 0.007 |
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".