Forecasting Air Pollution using a Modified Compositional Learning Approach
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
Major air pollutants, especially fine particles PM2.5, are generally associated with adverse health effects, including cardiac and respiratory morbidity. The aim of this paper is to find the best combination of machine learning techniques to forecast the Air Quality Index (AQI) using the Beijing air quality datasets. The dataset consists [among other] of six air pollutant attributes - PM2.5, PM10, SO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> , NO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> , CO and O <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf> that are considered important factors in calculating the Air Quality Index. Our initial results showed that Linear Regression model is not adequate in predicting and forecasting the air pollutants. Random Forest and Random Committee models performed better in terms of MAE and RMSE values compared to Linear Regression. Furthermore, it was noticed that Random Forest performs better in terms of accuracy for certain features but not all while Random Committee performs better in other set of features. This shows that using a "single" machine learning approach to predict or forecast the entire features set may not give the best accuracy. So, as a result, a modified compositional learning model with disentanglement using optimized hyperparameters and search space was designed. The results of this novel network show a marked improvement (3.34% to 78%) in terms of MAE and RMSE values when compared to Random Forest and Random Committee.
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) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.002 | 0.002 |
| 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 it