Kaggle数据挖掘竞赛冠军及优胜者代码汇总

原文:http://www.chioka.in/kaggle-competition-solutions/

We learn more from code, and from great code. Not necessarily always the 1st ranking solution, because we also learn what makes a stellar and just a good solution. I will post solutions I came upon so we can all learn to become better!

I collected the following source code and interesting discussions from the Kaggle held competitions for learning purposes. Not all competitions are listed because I am only manually collecting them, also some competitions are not listed due to no one sharing. I will add more as time goes by. Thank you.

Algorithmic Trading Challenge

  • Solution whitepaper.
  • Solution thread.

Allstate Purchase Prediction Challenge

  • Rank 2 solution code by Alessandro Mariani.
  • Rank 10 solution code by B1aine.
  • Rank 36 solution code by Hiroyuki.
  • Rank 159 solution code by MrCanard.
  • Solution thread.

Amazon.com – Employee Access Challenge

  • Rank 1 solution code by Paul Duan and Benjamin Solecki team.
  • Rank 1 solution Q&A by Paul Duan.
  • Rank 2 solution code by Owen Zhang.
  • Rank 3 solution code by Dmitry & Leustagos.
  • Rank 289 solution code by Foxtrot with original blog post here.
  • Solution thread.

AMS 2013-2014 Solar Energy Prediction Contest

  • Rank 1 solution code and description by Leustagos team.
  • Rank 2 solution code and description by Toulouse.
  • Rank 3 solution code 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 and here.

Belkin Energy Disaggregation Competition

  • Solution thread.

Challenges in Representation Learning: Facial Expression Recognition Challenge

  • Rank 1 solution code and description by Charlie Tang.
  • Rank 3 solution description by Maxim Milakov.
  • Solution thread.

Challenges in Representation Learning: The Black Box Learning Challenge

  • Rank 1 solution description by David Thaler.
  • Rank 2 solution code and description by sayit.

Challenges in Representation Learning: Multi-modal Learning

  • Rank 1 solution by MMDL.
  • Solution thread.

Detecting Insults in Social Commentary

  • Rank 1 solution description and code by Vivek Sharma.
  • Rank 2 solution 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.

EMI Music Data Science Hackathon

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

Galaxy Zoo – The Galaxy Challenge

  • Rank 1 solution code and description 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 description by hxu.
  • Rank 58 solution code and description by yr.
  • Solution thread.

Global Energy Forecasting Competition 2012 – Wind Forecasting

  • Rank 1 solution by Leustagos.
  • Solution thread here.

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

  • Rank 1 solution with code and description by Team Algorithm, Github link to code here.

KDD Cup 2013 – Author Disambiguation Challenge (Track 2)

  • Rank 1 solution with code and description by Team Algorithm, Github link to code here.
  • Rank 2 solution by SmallData Team.
  • Rank 3 solution by hustmonk.
  • Rank 4 solution by Ben S.
  • Solution thread.

Large Scale Hierarchical Text Classification

  • Rank 1 solution code and description by anttip.
  • Rank 3 solution code and description by nagadomi.
  • Solution thread one.
  • Solution thread two.

Loan Default Prediction – Imperial College London

  • Rank 2 solution and description by HelloWorld.
  • Rank 12 solution and description by David McGarry.
  • Solution thread.

Merck Molecular Activity Challenge

  • Ideas sharing discussion thread.
  • Preprocessing techniques discussion thread.

MLSP 2013 Bird Classification Challenge

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

Observing the Dark World

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

PAKDD 2014 – ASUS Malfunctional Components Prediction

  • Yet another solution thread.
  • Solution thread.

Personalize Expedia Hotel Searches – ICDM 2013

  • Presentation paper/slides for ICDM 2013.
  • Solution thread.

Predicting a Biological Response

  • Rank 6 solution by Shea Parkes & Neil Schneider team.
  • Rank 17 solution of Ensemble of RandomForests, GradientBoostingTrees and ExtraTreesRegressor by 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.

    Predicting Closed Questions on Stack Overflow

    • Rank 10 solution by Marco Lui.
    • Rank 33 solution by Foxtrot.

    See Click Predict Fix

    • Solution thread.

    See Click Predict Fix – Hackathon

    • Solution thread.

    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)

    The Analytics Edge (15.071x)

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

    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.

    Walmart Recruiting – Store Sales Forecasting

    • Rank 1 solution code and description by David Thaler.
    • Rank 2 solution description by sriok.
    • Rank 3 solution code and description 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:(机器学习)