Record Linkage Methodology for the Social Data Linkage Environment at Statistics Canada
Bibliographic record
Abstract
ABSTRACTObjectivesThe objectives of this talk are to introduce Statistics Canada’s Social Data Linkage Environment (SDLE) and to explain the methodology behind the creation of the central depository and how both deterministic and probabilistic record linkage techniques are used to maintain and expand the environment.ApproachWe will start with a brief overview of the SDLE and then continue with a discussion of how both deterministic linkages and probabilistic linkages (using Statistic Canada’s generalized record linkage software, G-Link) have been combined to create and maintain a very large central depository, which can in turn be linked to virtually any social data source for the ultimate end goal of analysis.ResultsAlthough Canada has a population of about 36 million people, the central depository contains some 300 million records to represent them, due to multiple addresses, names, etc. Although this allows for a significant reduction in missing links, it raises the spectre of additional false positive matches and has added computational complexity which we have had to overcome.ConclusionThe combination of deterministic and probabilistic record linkage strategies has been effective in creating the central depository for the SDLE. As more and more data are linked to the environment and we continue to refine our methodology, we can now move on to the ultimate goal of the SDLE, which is to analyze this vast wealth of linked data.
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 | 0.031 | 0.094 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.009 | 0.024 |
| Science and technology studies | 0.005 | 0.002 |
| Scholarly communication | 0.007 | 0.003 |
| Open science | 0.004 | 0.005 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.036 | 0.008 |
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".