Development of a new disease activity index for Systemic Sclerosis using traditional and machine learning techniques
Bibliographic record
Abstract
Scleroderma is an auto-immune disease characterized by thickened and hardened skin. Similarly to other rheumatic diseases such as Lupus, disease activity in Scleroderma greatly fluctuates over time. Developing a new disease activity index for Scleroderma is a necessary step before undertaking clinical trials to test different treatments and would provide a conceptual structure to approach this poorly-defined disease. In this thesis, we first apply statistical methods traditionally used to develop disease activity indices such as Factor Analysis, Principal Component Analysis and Multiple Linear Regression. We then compare these approaches to more modern statistical learning approaches such as Ridge and Lasso Regression, Principal Component Regression, Partial Least Squares Regression and Regression trees in the context of disease activity index construction and validation. We assess the predictive ability of the traditional and new methods using data from the Canadian Scleroderma Research Group (CSRG) registry. The methods are first assessed and compared by limiting the diagnostic criteria to those that are included in the commonly used Scleroderma Disease Activity Score (SDAS, Valentini et al., 2001 and 2003). In our work, we found that the SDAS does not predict physician global assessment of activity very well for patients in the CSRG registry. There are important discrepancies between the performance and generability of the index as reported by Valentini et al. and the results of our analyses. Thus, we conclude with the development of a new disease activity index for Scleroderma using the methods that previously showed good properties and a wider class of predictors. In summary, we found that the Lasso Regression approach outperforms other unsupervised and supervised learning techniques for predicting our outcome variables. It automatically selects good predictors and yields accurate prediction models, both in the context of the original SDAS and the new in
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 | 0.007 | 0.012 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 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".