Unsupervised learning approach for benchmark models to identify construction projects with high accident risk levels
Bibliographic record
Abstract
Purpose The construction sector is highly prone to accidents, traditionally assessed using subjective qualitative measurements. To enhance the allocation of risk management resources and identify high-risk projects during pre-construction, an objective and quantitative approach is necessary. This study introduces a three-step clustering methodology to quantitatively evaluate accident risk levels in construction projects. Design/methodology/approach In the first step, accident and total construction revenue by project were collected to calculate accident probabilities. In the second step, accident probabilities were calculated by project type using the data collected in the first step. After that, benchmark models were suggested using clustering methods to identify high-risk project types for risk management. Before suggesting the benchmark models, an uncertainty analysis was conducted due to the limited amount of data. In the third step, the suggested benchmark models were validated for accuracy. Findings The results categorized risk levels for fatalities and injuries into four distinct groups. Validation through ordinal logistic regression demonstrated high explanatory power, with fatality risk levels ranging from 79.9 to 100% and injury risk levels from 90.3 to 100%. Originality/value This benchmark model facilitates effective comparisons and analyses across various construction sectors and countries, offering a robust quantitative standard for risk management. By identifying high-risk projects such as “Dam,” this methodology enables better resource allocation during the pre-construction phase, thereby improving overall safety management in the construction industry and providing a basis for legislative applications.
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 | 0.003 | 0.010 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".