A density-based data cleaning approach for deduplication with data consistency and accuracy
Bibliographic record
Abstract
Data cleaning is a critical part of the data transformation stage in data warehousing where the extracted data from relational databases are usually unclean. This may affect critical tasks in different organizations such as data analysis and decision making. Current techniques of data cleaning generally deal with one or two quality aspects. The techniques assume the availability of master data, or that users are involved in data cleaning such as manually placing confidence scores that represent the correctness of the values of data. In this paper, we present a uniform framework and algorithms to integrate data deduplication with inconsistent data repairing and discovering of the accurate values in data. We utilize the embedded density information in data to fix errors based on data density where tuples that are close to each other are packed together. We present a weight model to assign confidence scores that are based on the density of data. The assignments are automated and no user is involved in the process. We consider the inconsistent data in terms of violations with respect to a set of functional dependencies (FDs), as these violations are common in practice. We present a cost model for data repairing that is based on the weight model. We experimentally verify the quality and the scalability of our algorithms. We use synthetic and real datasets.
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 | 0.006 | 0.031 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.004 | 0.005 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.003 | 0.006 |
| Open science | 0.005 | 0.005 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 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".