Predicting the structural condition of individual sanitary sewer pipes with random forests
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
Closed-circuit television inspections of sewer condition deterioration as required for proactive management are expensive and hence limited to portions of a sewer network. The data mining approach presented herein is shown capable of unlocking information contained within inspection records and enhances existing pipe inspection practices currently used in the wastewater industry. Predictive models developed using the random forests algorithm are found capable of predicting individual sewer pipe condition so that uninspected pipes in a sewer network with the greatest likelihood of being in a structurally defective condition state are identified for future rounds of inspection. Complications posed by imbalance between classes common within inspection datasets are overcome by first establishing the classification task in a binary format (where pipes are in either good or bad structural condition) and then using the receiver-operating characteristic (ROC) curve to establish alternative cutoffs for the predicted class probability. The random forests algorithm achieved a stratified test set false negative rate of 18%, false positive rate of 27% and an excellent area under the ROC curve of 0.81 in a case study application to the City of Guelph, Ontario, Canada. The novel inclusion of condition information of pipes attached at either the upstream or downstream manholes of an individual pipe enhances the predictive power for bad pipes representing the minority class of interest (reducing the false negative rate to 11%, reducing the false positive rate to 25% and increasing the area under the ROC curve to 0.85). An area under the ROC curve >0.80 indicates random forests are an “excellent” choice for predicting the condition of individual pipes in a sewer network.
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 | 0.000 | 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 | 0.000 |
| Open science | 0.000 | 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 it