Skip to content

README.md

Latest commit

 

History

History
30 lines (21 loc) · 1.15 KB

README.md

File metadata and controls

30 lines (21 loc) · 1.15 KB

pnn module

This module has the following components:

  • pnn.nn: Neural network architectures and functionality relating to file input/output, training, testing, estimation, etc.

  • pnn.output: Scientific outputs, such as drawing figures, writing out tables, etc.

  • pnn.aggregate: Helper functions for aggregating results, e.g. calculating MdSA and other metrics for each scenario-architecture combination.

  • pnn.constants: Constants used elsewhere in the code, such as default file locations and parameters. Provides the Parameter dataclass which ensures consistent nomenclature, units, colours, etc. for scenarios, IOPs, uncertainty types, and metrics.

  • pnn.data: Reading and pre-processing input data (original and split) as well as re-scaling of IOPs.

  • pnn.maps: Applying neutral network modules to PRISMA scenes and plotting the results.

  • pnn.metrics: Calculating metrics like MdSA, SSPB, R², coverage, etc.

  • pnn.modeloutput: Reading and pre-processing model outputs (IOP estimates).

  • pnn.recalibration: Neural network recalibration; fitting and applying recalibration functions, calculating calibration curves.