MATLAB API
The MATLAB implementation provides a reference interface for PME-toolkit workflows.
Setup
Add the source folder to the MATLAB path:
addpath(genpath("matlab/src"));
Running a case
A dimensionality reduction workflow can be executed from a JSON configuration file:
out = pme.run_case("tests/cases/test_glider.json");
Backmapping
Backmapping can be performed using a dedicated configuration:
out = pme.backmapping("tests/cases/test_glider_back.json");
Output
The returned structure out typically includes:
- reduced coordinates
- variance information
- reconstruction metrics
- model data
Exact contents depend on the configuration.
Wrappers
Repository-level wrappers are available:
run_pme.mrun_back.m
These provide a simplified interface to the same JSON-driven workflow:
run_pme("tests/cases/test_glider.json")
run_back("tests/cases/test_glider_back.json")
Notes
- MATLAB and Python implementations follow the same configuration structure
- JSON files define the full workflow
- MATLAB is useful for validation and comparison with the reference implementation