Using Prior Information on Parameters to Eliminate Dependence on Initial Values in Fitting Coxian Phase Type Distributions to Length of Stay Data in Healthcare Settings
Bibliographic record
Abstract
Background: Modeling length of stay (LOS) data in healthcare settings using Coxian phase type (PH) distributions is becoming increasingly popular. However, dependence on initial values is a persistent difficulty in parameter estimations. This paper explores the utility of prior information on the parameters to address this difficulty. Methods: Maximum likelihood methods were used to estimate parameters of PH distributions that best fit simulated datasets with various sample sizes arising from PH distributions of various numbers of phases and parameters, using randomly generated initial values. Estimated values for the parameters resulting from different initial values were compared to the known values to assess the extent to which estimates depend on initial values; the impacts of sample sizes, existence of prior information, as well as the number of parameters with prior information were assessed. Results: Without prior information, parameter estimates depend on initial values for all PH distributions and all sample sizes. Prior information on one or more parameters led to more concentrated estimates, with higher number of parameters with prior information or larger sample sizes leading to more concentrated estimates. For example, with a sample size of 500, the estimates for a parameter with known value of 0.706 without prior information had a wide range of 1.523; using prior information for two parameters narrowed that range down to 0.156. For 3-phase PH distributions, prior information on 3 parameters appeared to be sufficient to eliminate dependence on initial values, even for small sample sizes. For 4-phase PH distributions, prior information on 5 parameters and a moderate sample size were needed to eliminate such dependence. Conclusions: Combination of prior information on parameters and sufficient sample sizes can eliminate dependence on initial values in fitting PH distributions to LOS 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 | 0.046 | 0.204 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.002 | 0.004 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.004 |
| Insufficient payload (model declined to judge) | 0.004 | 0.000 |
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".