/lstm-da-analysis

Analysis of data assimilation results for long short-term memory netoworks

Primary LanguageJupyter Notebook

Analyzing results from Data Assililation with LSTM

Main analysis done with Jupyter Notebook:

  • assimilation_analysis_jmframe.ipynb data_assimilation_performance_difference autoregression_performance_difference

./data/

To run the notebook there should be three pickle files with results here:

  • simulation/.../test_results.p
  • data_assimilation/.../test_results.p
  • autoregression/.../test_results.p

There also should be some specific files for the analysis:

  • camels_attributes_v2.0/
  • usgs_site_info.csv

./tools/

  • analysis_tools.py Has a bunch of functions used in the "assimilation_analysis_jframe.ipynb" workbook
  • corrstats.py Functions for calculating the statistical significant differences between two dependent or independent correlation coefficients
  • metrics.py Metrics taken from NeuralHydrology code
  • signatures.py Hydrologic signature calculations taken from NeuralHydrology code