Multiclass anomaly detection in imbalanced structural health monitoring data using convolutional neural network
Bibliographic record
Abstract
Abstract Structural health monitoring (SHM) system aims to monitor the in-service condition of civil infrastructures, incorporate proactive maintenance, and avoid potential safety risks. An SHM system involves the collection of large amounts of data and data transmission. However, due to the normal aging of sensors, exposure to outdoor weather conditions, accidental incidences, and various operational factors, sensors installed on civil infrastructures can get malfunctioned. A malfunctioned sensor induces significant multiclass anomalies in measured SHM data, requiring robust anomaly detection techniques as an essential data cleaning process. Moreover, civil infrastructure often has imbalanced anomaly data where most of the SHM data remain biased to a certain type of anomalies. This imbalanced time-series data causes significant challenges to the existing anomaly detection methods. Without proper data cleaning processes, the SHM technology does not provide useful insights even if advanced damage diagnostic techniques are applied. This paper proposes a hyperparameter-tuned convolutional neural network (CNN) for multiclass imbalanced anomaly detection (CNN-MIAD) modelling. The hyperparameters of the proposed model are tuned through a random search algorithm to optimize the performance. The effect of balancing the database is considered by augmenting the dataset. The proposed CNN-MIAD model is demonstrated with a multiclass time-series of anomaly data obtained from a real-life cable-stayed bridge under various cases of data imbalances. The study concludes that balancing the database with a time shift window to increase the database has generated the optimum results, with an overall accuracy of 97.74%.
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 | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| 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 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".