zooie / tagger Goto Github PK

@author: Vik Singh ([email protected])

A simple BOSS example for Yahoo! Hack Day NYC
A machine learned tagger trained on BOSS web / delicious data

Read this if you want to learn more and especially check out the caveats section if
you're planning to use this code for more practical purposes

http://zooie.wordpress.com/2009/10/09/build-an-automatic-tagger-in-200-lines-with-boss/


# Install libsvm

tar -xzvf libsvm-2.89.tar.gz
cd libsvm-2.89
make

cd ..

# Optional: Crawl fresh delicious data via BOSS (a previous crawl already included)

python crawl_delicious.py

# Generate a binary training set via two tags (pick from tags.txt)

python gen_training_test_set.py microsoft google

# Learn from the resulting training_data.txt and predict on test_data.txt

python autosvm.py training_set.txt test_set.txt

# Prints out the accuracy of the learner and saves model + prediction files in timestamped folder