Multivariate Data Cleaning and Classification
Bibliographic record
Abstract
Geological data sets often contain some errors. These errors might be linked to different issues like measurement devise limitations, measurement recording glitches or simply interpretation inconsistencies related to change in interpreter and/or interpretation concept. However, irrelevant of the reason for erroneous data, every geomodeler, geologist and/or engineer have to deal with data problem issues as otherwise he or she would waste days drawing wrong conclusions because the errors have not been first identified and excluded. In the literature one can find several methods for data cleaning, the most common being statistical methods (standard deviation, range, or clustering algorithms), data transformation, and duplicate elimination. Not all of the approaches are equally useful and effective. In this talk we present and further develop a validated multivariate cleaning methodology based on the Mahalanobis distance approach. This method is widely used cluster analysis and classification techniques; application to geological data is novel. After the methodology is explained in detail it is illustrated using the core well data from the Joslyn Project. In particular, it is shown on the Jolsyn example how to identify and analyze errors in datasets containing many variables, including bitumen grade and particle size distributions (psd’s) characterizing rock granulometry and specified by 22 bin sizes. Analysis is done on a by-facies basis. Furthermore, to avoid losing valuable information, for example, misclassified psd, geologically driven classification is developed. This classification incorporates multivariate geological information and allows assigning mislabeled data to the best suitable facies group or class. The proposed classification not only makes geological sense, but also makes further geological analysis more consistent and straightforward.
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 | 0.008 | 0.034 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.006 | 0.007 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.004 | 0.002 |
| Open science | 0.003 | 0.004 |
| Research integrity | 0.001 | 0.003 |
| Insufficient payload (model declined to judge) | 0.007 | 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".