Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2025
|
| Online Access: | https://doi.org/10.5281/zenodo.17847453 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866902337110409216 |
|---|---|
| author | xuc1998 |
| author_facet | xuc1998 |
| contents | <p>First public release of the PATPEMS MATLAB package.</p> <p>This version contains:</p> <p>Core implementation of the Parallel Adaptive Transition Particle Evolution Metropolis Sequential Monte Carlo (PATPEMS) algorithm, an enhanced framework that emphasizes robustness and practicality through three refinements:</p> <p>Parallelism (P) exploits concurrency in both the reweighting and moving steps, yielding clear wall-clock speedups for likelihood-intensive models.</p> <p>Adaptive transition (AT) designs the sequence of intermediate distributions so that increments are aligned with weight stability, thereby constraining stage growth and improving sample efficiency.</p> <p>Flexible scheduling of move operators (PEM) enables problem-tailored ordering and repetition of RWM, crossover, and DE-MH operators, coupled with two boundary handlers, to preserve reversibility and reduce wasted proposals.</p> <p>Installation and usage instructions.</p> <p>Four fully configured case-study examples:</p> <p>Example 1: 2-D 20-mode Gaussian mixture (multi-modal benchmark).</p> <p>Example 2: 100-D bimodal Gaussian (high-dimensional benchmark).</p> <p>Example 3: Synthetic CoLM calibration (6-parameter, no model–data mismatch).</p> <p>Example 4: Real-world CoLM calibration at the Arou site.</p> <p>Basic plotting and diagnostic utilities for ESS/rCESS, log-evidence, acceptance rates, and marginal/posterior visualisation.</p> <p>Requirements:</p> <p>MATLAB R2020b or later (tested on R2022a).</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_17847453 |
| institution | Zenodo |
| language | |
| publishDate | 2025 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | xuc1998/PATPEMS-Algorithm: PATPEMS v1.0.0 – MATLAB implementation and case-study examples xuc1998 <p>First public release of the PATPEMS MATLAB package.</p> <p>This version contains:</p> <p>Core implementation of the Parallel Adaptive Transition Particle Evolution Metropolis Sequential Monte Carlo (PATPEMS) algorithm, an enhanced framework that emphasizes robustness and practicality through three refinements:</p> <p>Parallelism (P) exploits concurrency in both the reweighting and moving steps, yielding clear wall-clock speedups for likelihood-intensive models.</p> <p>Adaptive transition (AT) designs the sequence of intermediate distributions so that increments are aligned with weight stability, thereby constraining stage growth and improving sample efficiency.</p> <p>Flexible scheduling of move operators (PEM) enables problem-tailored ordering and repetition of RWM, crossover, and DE-MH operators, coupled with two boundary handlers, to preserve reversibility and reduce wasted proposals.</p> <p>Installation and usage instructions.</p> <p>Four fully configured case-study examples:</p> <p>Example 1: 2-D 20-mode Gaussian mixture (multi-modal benchmark).</p> <p>Example 2: 100-D bimodal Gaussian (high-dimensional benchmark).</p> <p>Example 3: Synthetic CoLM calibration (6-parameter, no model–data mismatch).</p> <p>Example 4: Real-world CoLM calibration at the Arou site.</p> <p>Basic plotting and diagnostic utilities for ESS/rCESS, log-evidence, acceptance rates, and marginal/posterior visualisation.</p> <p>Requirements:</p> <p>MATLAB R2020b or later (tested on R2022a).</p> |
| title | xuc1998/PATPEMS-Algorithm: PATPEMS v1.0.0 – MATLAB implementation and case-study examples |
| url | https://doi.org/10.5281/zenodo.17847453 |