Skip to content

ckt624/Kaggle_CrowdFlower

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Kaggle_CrowdFlower

1st Place Solution for Search Results Relevance Competition on Kaggle

The best single model we have obtained during the competition was an XGBoost model with linear booster of Public LB score 0.69322 and Private LB score 0.70768. Our final winning submission was a median ensemble of 35 best Public LB submissions. This submission scored 0.70807 on Public LB and 0.72189 on Private LB.

What's New

FlowChart

FlowChart

Documentation

See ./Doc/Kaggle_CrowdFlower_ChenglongChen.pdf for documentation.

Instruction

  • download data from the competition website and put all the data into folder ./Data.
  • run python ./Code/Feat/run_all.py to generate features. This will take a few hours.
  • run python ./Code/Model/generate_best_single_model.py to generate best single model submission. In our experience, it only takes a few trials to generate model of best performance or similar performance. See the training log in ./Output/Log/[Pre@solution]_[Feat@svd100_and_bow_Jun27]_[Model@reg_xgb_linear]_hyperopt.log for example.
  • run python ./Code/Model/generate_model_library.py to generate model library. This is quite time consuming. But you don't have to wait for this script to finish: you can run the next step once you have some models trained.
  • run python ./Code/Model/generate_ensemble_submission.py to generate submission via ensemble selection.
  • if you don't want to run the code, just submit the file in ./Output/Subm.

About

1st Place Solution for Search Results Relevance Competition on Kaggle (https://www.kaggle.com/c/crowdflower-search-relevance)

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C++ 66.2%
  • Python 26.1%
  • Logos 5.5%
  • Yacc 1.6%
  • Perl 0.3%
  • TeX 0.1%
  • Other 0.2%