Skip to content
C++ code for "A Faster Drop-in Implementation for Leaf-wise Exact Greedy Induction of Decision Tree Using Pre-sorted Deque"
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
build
include
test
tools
.gitignore
Makefile
README.md

README.md

Leaf-wise Induction of Decision Tree with Presorted Deque

This is the proof-of-concept demo code for reproducing experiments in the arXiv note "A Faster Drop-in Implementation for Leaf-wise Exact Greedy Induction of Decision Tree Using Pre-sorted Deque" (https://arxiv.org/abs/1712.06989).

Prepare sample data

  • download data from HIGGS and uncompress gz file.
  • create training data head -1000000 HIGGS.csv > higgs-train-1m.csv
  • create testing data tail -50000 HIGGS.csv > higgs-test.csv

Compile and test

$ make
$ OMP_NUM_THREADS=28 ./build/test_dt higgs-train-1m.csv higgs-test.csv
tree induction time: 1.217055 seconds
training time: 2.237200 seconds
nleafs: 2216 
test accuracy: 0.710

Other tests on synthetic data

$ OMP_NUM_THREADS=8 ./build/test_dt
tree induction time: 2.715179 seconds
training time: 4.514815 seconds
nleafs: 33
test accuracy: 0.973

All rights reserved. Jianbo Ye

You can’t perform that action at this time.