Efficient detection of data entry errors in large-scale public health surveys: an unsupervised machine learning approach
Bibliographic record
Abstract
Data entry errors in large-scale public health surveys can undermine the effectiveness of data-driven interventions. Therefore, identifying these data entry errors is crucial for public health experts. In large-scale public health surveys, manually verifying the accuracy of every data point by domain experts is nearly impossible. This study evaluates unsupervised machine learning algorithms for detecting these errors, focusing on the 'weight' parameter in the Annual Health Survey (AHS) dataset. The AHS, conducted by the Ministry of Health and Family Welfare, Government of India, in collaboration with the Registrar General of India, is a large-scale, stratified, household-level survey targeting maternal and child health across nine states in India. The dataset is freely available on the Open Government Data (OGD) Platform of India for public health research. In this study, five algorithms—DBSCAN, K-Means, Gaussian Mixture Model (GMM), Isolation Forest (IF), and One-Class SVM (1C-SVM) were applied to detect erroneous data entries. The evaluation process involved comprehensive preprocessing and feature engineering to optimize detection capabilities. Performance metrics such as precision, recall, accuracy, false anomaly, and missed anomaly rates were used to assess each algorithm. Among these, DBSCAN demonstrated superior performance, achieving a recall of 94.7% and a precision of 81.9%, making it highly effective for this task. The findings underscore the potential of unsupervised machine learning in automating the detection of data entry errors, thereby improving the integrity of public health data. This research contributes to the advancement of precision public health, supporting more accurate and reliable evidence-based decision-making and policy formulation.
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 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 | 0.034 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.003 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.003 |
| 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 teacher head, 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".