Code for Python digit recognizer starter comptetition. Checking knowledge in ML (http://www.kaggle.com/c/digit-recognizer)
Pull request Compare This branch is 1 commit behind dzhibas:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
benchmark
data
py-knn
py-rf
.gitignore
README.md

README.md

kaggle-digit-recognizer

Code for Python digit recognizer starter comptetition. http://www.kaggle.com/c/digit-recognizer

benchmark kNN k=10, cover_tree gives 0.96557 accuracy benchmark rf 1000 trees gives 0.96829 accuracy

knnpy-knn/experiment2 gives same accuracy as benchmark, but if we reduce dimmensions with PCA to lets say ~100 components kNN performs better than rf benchmark.

plotted explained variance of PCA helps to choose optimal component count to keep 90% variance

py-knn

For kNN experiments

py-rf

For random forests experiments

libraries used

  • numpy
  • scipy
  • sklearn (scikit-learn)
  • matplotlib