Skip to content

vbugaevskii/IDAO-2019-muon-id

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

11 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

IDAO-2019-muon-id

This repository is the solution for the Online Round of IDAO-2019 ML-contest of team AR_U_KIDDIN_MI.

Team members:

Baseline solution from contest's organizers can be found here.

Track 1

Notebook Description
Features-FOI.ipynb Notebook contains engineering of advanced features based on FOI features.
Track1-MLPClassifier-fit.ipynb Notebook contains feature engineering and fitting processes for an ensemble of Multilayer Perceptron Classifiers.
Track1-MLPClassifier-predict-public.ipynb Notebook contains prediction process for an ensemble of Multilayer Perceptron Classifiers for public test dataset.
Track1-MLPClassifier-predict-private.ipynb Notebook contains prediction process for an ensemble of Multilayer Perceptron Classifiers for private test dataset.
Track1-CatBoostClassifier-fit.ipynb Notebook contains feature engineering and fitting processes for an ensemble of CatBoostClassifiers.
Track1-CatBoostClassifier-predict-public.ipynb Notebook contains prediction process for an ensemble of CatBoostClassifiers for public test dataset.
Track1-CatBoostClassifier-predict-private.ipynb Notebook contains prediction process for an ensemble of CatBoostClassifiers for private test dataset.
Track1-PredictionsMerge.ipynb Notebook contains a weighted merge of predictions done by ensembles of MLPClassifiers and CatBoostClassifiers.

More information about HEP MLPClassifier can be found here.

Track 2

Features-FOI.ipynb contains engineering of advanced features based on FOI features.

Track2-CatBoostClassifier.ipynb contains feature engineering and fitting model processes.

Track2-CPP contains C++ code for submission.

Note: Final submission for track 2 is the advanced baseline submission. CatBoostClassfier's parameters:

{
  "iterations": 1700,
  "max_depth": 5
}

Leaderboard

public private
Track 1 7613.14 7784.81
Track 2 7253.11 7567.33

Our team took the 10th place on the private leaderboard according to both tracks.

Note: Original score is multiplied by 10 000.

About

No description or website provided.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published