【转】Kaggle 机器学习竞赛冠军及优胜者的源代码汇总

Kaggle比赛源代码和讨论的收集整理。

Algorithmic Trading Challenge25

  • Solution whitepaper26.
  • Solution thread19.
回到顶部

Allstate Purchase Prediction Challenge3

  • Rank 2 solution code21 by Alessandro Mariani.
  • Rank 10 solution code5 by B1aine.
  • Rank 36 solution cod1e by Hiroyuki.
  • Rank 159 solution code by MrCanard.
  • Solution thread.
回到顶部

Amazon.com – Employee Access Challenge6

  • Rank 1 solution code24 by Paul Duan and Benjamin Solecki team.
  • Rank 1 solution Q&A5 by Paul Duan.
  • Rank 2 solution code1 by Owen Zhang.
  • Rank 3 solution code3 by Dmitry & Leustagos.
  • Rank 289 solution code by Foxtrot with original blog post here.
  • Solution thread.
回到顶部

AMS 2013-2014 Solar Energy Prediction Contest2

  • Rank 1 solution code5 and description4 by Leustagos team.
  • Rank 2 solution code and description by Toulouse.
  • Rank 3 solution code1 and description by Owen Zhang.
  • Rank 4 solution escription by Peter Prettenhofer.
  • Rank 5 solution description by Domcastro.
  • Rank 58 solution code and description by Davit.
  • Solution thread here.
  • Ridge Regression starter code with MAE about 2.2M by Alec Radford, original thread here.
  • Improved starter code by Foxtrot.
  • Baseline code with MAE about 2.6M using Catmull-Rom Spline interpolation, also available in R here andhere.
回到顶部

Belkin Energy Disaggregation Competition1

  • Solution thread.
回到顶部

Challenges in Representation Learning: Facial Expression Recognition Challenge4

  • Rank 1 solution code6 and description2 by Charlie Tang.
  • Rank 3 solution description3 by Maxim Milakov.
  • Solution thread.
回到顶部

Challenges in Representation Learning: The Black Box Learning Challenge1

  • Rank 1 solution description1 by David Thaler.
  • Rank 2 solution code and description by sayit.
回到顶部

Challenges in Representation Learning: Multi-modal Learning2

  • Rank 1 solution1 by MMDL.
  • Solution thread.
回到顶部

Detecting Insults in Social Commentary

  • Rank 1 solution description4 and code by Vivek Sharma.
  • Rank 2 solution1 by tuzzeg.
  • Rank 3 solution description Andrei Olariu.
  • Rank 4 solution by Chris Brew.
  • Rank 5 solution description by Yasser Tabandeh.
  • Rank 6 solution by Andreas Mueller, code available here.
  • Rank 8 solution description by Steve Poulson.
  • Solution thread.
回到顶部

EMI Music Data Science Hackathon

  • Rank 4 solution description1 by Steffen Rindle.
  • Rank 18 solution code and description by Vlad Gusev.
  • Rank 34 solution code and description by zenog.
  • Solution thread.
回到顶部

Galaxy Zoo – The Galaxy Challenge

  • Rank 1 solution code2 and description1 by Sander Dieleman.
  • Rank 2 solution code and description by Maxim Milakov.
  • Rank 3 solution code and description by tund.
  • Rank 5 solution code and description by Julian de Wit.
  • Rank 9 solution code and description by Soumith Chintala.
  • Rank 13 solution code and description by Xiaoxiang Zhang.
  • Rank 28 solution code and description by utdiscant.
  • Rank 38 solution code and description by sugi.
  • Rank 57 solution code and description1 by hxu.
  • Rank 58 solution code and description by yr.
  • Solution thread.
回到顶部

Global Energy Forecasting Competition 2012 – Wind Forecasting

  • Rank 1 solution by Leustagos.
  • Solution thread here1.
回到顶部

KDD Cup 2013 – Author-Paper Identification Challenge (Track 1)2

  • Rank 1 solution with code and description4 by Team Algorithm, Github link to code here1.
回到顶部

KDD Cup 2013 – Author Disambiguation Challenge (Track 2)1

  • Rank 1 solution with code and description4 by Team Algorithm, Github link to code here1.
  • Rank 2 solution1 by SmallData Team.
  • Rank 3 solution1 by hustmonk.
  • Rank 4 solution1 by Ben S.
  • Solution thread1.
回到顶部

Large Scale Hierarchical Text Classification4

  • Rank 1 solution code and description7 by anttip.
  • Rank 3 solution code2 and description2 by nagadomi.
  • Solution thread one3.
  • Solution thread two2.
回到顶部

Loan Default Prediction – Imperial College London

  • Rank 2 solution and description1 by HelloWorld.
  • Rank 12 solution and description by David McGarry.
  • Solution thread.
回到顶部

Merck Molecular Activity Challenge1

  • Ideas sharing discussion thread.
  • Preprocessing techniques discussion thread.
回到顶部

MLSP 2013 Bird Classification Challenge

  • Rank 1 solution code3 and description by beluga.
  • Rank 2 solution code1 and description by Herbal Candy (W and thomeou).
  • Rank 3 solution description by Anil Thomas.
  • Rank 4 solution description by Maxim Milakov.
  • Solution thread.
回到顶部

Observing the Dark World

  • Rank 2 solution by Iain Murray, code available here.
回到顶部

PAKDD 2014 – ASUS Malfunctional Components Prediction

  • Yet another solution thread1.
  • Solution thread2.
回到顶部

Personalize Expedia Hotel Searches – ICDM 2013

  • Presentation paper/slides1 for ICDM 2013.
  • Solution thread1.
回到顶部

Predicting a Biological Response1

  • Rank 6 solution by Shea Parkes & Neil Schneider team.
  • Rank 17 solution of Ensemble of RandomForests, GradientBoostingTrees and ExtraTreesRegressorby Emanuele Olivetti.
  • Another solution code by Oblique Random Forest (oRF) by Shea Parkes & Neil Schneider team.
  • The code of my best submission thread. Talks about Multi-core training Oblique Random Forests, and Stacking.
  • Question about the process of ensemble learning thread. Talks about applying ensembles in practice, and how can problems arise and how to deal with them.
回到顶部

Predicting Closed Questions on Stack Overflow

  • Rank 10 solution by Marco Lui.
  • Rank 33 solution by Foxtrot.
回到顶部

See Click Predict Fix1

  • Solution thread1.
回到顶部

See Click Predict Fix – Hackathon1

  • Solution thread.
回到顶部

StumbleUpon Evergreen Classification Challenge

  • Benchmark beater 1.
  • Benchmark beater 2.
  • Benchmark beater 3.
  • Solution thread.
  • My own solution, which is a good example of what is overfitting. (Public rank: 57, Private rank: 291)
回到顶部

[The Analytics Edge (15.071x)](The%20Analytics Edge (15.071x))

  • Rank 17 solution code and description by Foxtrot.
  • Solution thread.
回到顶部

The Marinexplore and Cornell University Whale Detection Challenge

  • Rank 1 solution by Nick Kridler.
  • Rank 7 solution by Gilles Louppe and Peter Prettenhofer team.
  • Rank 8 solution by Sander Dieleman.
  • Rank 56 solution by Sudeep Juvekar.
  • Solution discussion thread.
  • Mean spectogram thread.
  • Official interview from the Marinexplorer and Cornell at Kaggle.
回到顶部

Walmart Recruiting – Store Sales Forecasting1

  • Rank 1 solution code5 and description by David Thaler.
  • Rank 2 solution description1 by sriok.
  • Rank 3 solution code and description1 by James King.
  • Rank 5 solution description by ACS69.
  • Rank 6 solution description by T. Henry.
  • Rank 8 solution description by BreakfastPirate.
  • Rank 9 solution description by Neil Summers.
  • Rank 10 solution description by Gilberto Titericz Junior.
  • Rank 11 solution description by citynight.
  • Rank 16 solution code and description by yr.
  • Rank 29 solution code and description by Mike Kim.
  • Rank 30 solution description by dkay.
  • Solution thread.

Thank you Foxtrot, James Petterson, Ben S for providing some of the links and solutions above.

你可能感兴趣的:(Python)