Application of Bayesian variable selection methods and shrinkage priors to epidemiological data
Bibliographic record
Abstract
First developed in the 1970s, variable selection plays a vital role in selecting the correct inclusion of variables in a prediction model. From forward selection to penalized regression, a great number of approaches have been constructed within the frequentist framework. More recently, Bayesian variable selection techniques have been developed and gained popularity because of their ability to learn from prior knowledge and construct credible intervals for the parameters without additional computations. However, the application of the Bayesian techniques on data with missingness or spatial information have been limited and the use of customizable programs such as JAGS and RStan are required.In this thesis, we provide a comprehensive review of some commonly-used variable selection methods, especially the Bayesian priors, and the comparisons of these methods from past literature. We then apply the regularized Horseshoe prior to two epidemiological datasets to investigate: (1) the random effect of health care regions and social-material deprivation on the treatment decision for patients with aortic stenosis, and (2) the correlations between epidemiological factors and HIV status obtained from a HIV self-testing arm survey. The shrinkage prior is first applied to the AS treatment dataset from the Institut national de santé publique du Québec (INSPQ). Spatial information is contained in postal codes and treated as random effects with 0-1 adjacency matrix. We then analyze the HIV data obtained from a quasi-randomized control trial to explore the risk factors of human immunodeficiency virus (HIV) status. With missing values, we apply a Bayesian hierarchical model for imputation and the shrinkage prior to drop irrelevant predictors. At last, we provide our code and offer ideas for future research
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 | 0.046 | 0.109 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.003 | 0.003 |
| Bibliometrics | 0.004 | 0.005 |
| Science and technology studies | 0.001 | 0.003 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.002 | 0.006 |
| Insufficient payload (model declined to judge) | 0.003 | 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".