Skip to content

HSE-LAMBDA/SHAD-2020-muon-id

 
 

Repository files navigation

Baseline solutions for the SHAD muon identification competition

Baseline - simple.ipynb contains a naive xgboost solution and creates a model file for Track 2

Baseline - advanced.ipynb contains a primer on jagged arrays processing and creates a model file for Track 2

track_2_baseline_advanced_cpp contains an example submission for Track 2 with catboost and jagged arrays in C++. Run make all to build and the baseline binary will be accepting data as .csv (not gzipped) at the standard input and writing predictions to the standard output. Run make compute_features to build an utility binary that preprocesses data and outputs the computed features.

track_2_baseline_simple_python contains an example submission for Track 2 with xgboost and python. Quite slow to read the data and doesn't have the jagged arrays.

scoring.py contains an implementation of the quality metric

utils.py contains data reading and jagged arrays processing utilities

Evaluate advanced baseline.ipynb compares Python and C++ predictions for Track 2

Packages

No packages published

Languages

  • Jupyter Notebook 85.5%
  • C++ 14.2%
  • Other 0.3%