Wikipedia LSTM

This repository describes the inmplementation of MLSL specifically to Wikipedia data.

See project page for the overall project, papers, and data.

The implementation is independent of the language of wikipedia being used.

Running Code

In order to run the code, currently data needs to be placed in results directory under wiki name. For example, for bgwiki, unzip the data into results as results/bgwiki/
Inside this wiki directory, there should be a file reduced_user_graph.json
Data for astwiki can be obtained from: https://drive.google.com/file/d/0B7XE3zJNvKQXQ09RcmJWN3A2V3M/view?usp=sharing
Then in the file graph_analysis.py, make sure that WIKINAME is same as the wiki being used
Run python graph_analysis.py to run MLSL
Results are stored in results/wikiname directory

Name		Name	Last commit message	Last commit date
Latest commit History 77 Commits
data		data
results		results
.gitignore		.gitignore
README.md		README.md
analysis.py		analysis.py
basics.py		basics.py
chdiff.py		chdiff.py
db.py		db.py
editdist.py		editdist.py
graph_analysis.py		graph_analysis.py
json_plus.py		json_plus.py
learning.py		learning.py
lstm.py		lstm.py
measure_significance.py		measure_significance.py
multi_LSTM.py		multi_LSTM.py
multi_layer_lstm_lstm.py		multi_layer_lstm_lstm.py
network.py		network.py
nn_base.py		nn_base.py
serializer.py		serializer.py
user_graph.py		user_graph.py
wikipedia.py		wikipedia.py