Skip to content
master
Switch branches/tags
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
lib
 
 
 
 
 
 
 
 
 
 
 
 

PAL

Pereference/Ranking Aggregation Library

All ideas and contributions are welcome. If you have any suggestions, you can contanct Pooya Moradi (po.moradi -at- ut.ac.ir) or Hamed Zamani (hamedzamani -at- acm.org). Please let us know if you found any problems/bugs in the code.

About

PAL is a ranking aggregation Ruby library developed in Intelligent Information Systems Labratory of School of ECE at University of Tehran.

Supported Methods

  • Weighted Kemeny Ranking
  • Borda Count
  • Copeland
  • Schulze

Getting started

At first you should clone the library.

git clone git@github.com:ut-iis/PAL.git

Now you should edit the sample config file according to your needs.

method: kemeny
out: /home/pooya/Projects/PAL/samples/output.dat
estimators:
    listnet:
      path: /home/pooya/Projects/PAL/samples/listnet.dat
      weight: 2
    ranknet:
      path: /home/pooya/Projects/PAL/samples/ranknet.dat
      weight: 3
    listmle:
      path: /home/pooya/Projects/PAL/samples/listmle.dat
      weight: 3
    svm:
      path: /home/pooya/Projects/PAL/samples/svm.dat
      weight: 1
evaluation:
    solution_file: /home/pooya/Projects/PAL/samples/solution.dat

Then in the rank.rb file you should address the config file and set if you want to evaluate the aggregation or not.

require './lib/pal.rb'

pal = PAL.new("/home/pooya/Projects/PAL/lib/inits/config.yml")
pal.run
#pal.evaluate("NDCG@10")

And finally:

ruby rank.rb

Citation

You can cite our paper published in ACM RecSys 2014:

@inproceedings{Zamani:2014,
  author    = {Zamani, Hamed and
               Shakery, Azadeh and
			   Moradi, Pooya},
  title     = {Regression and Learning to Rank Aggregation for User Engagement Evaluation},
  booktitle = {Proceedings of the 2014 Recommender Systems Challenge},
  series = {RecSysChallenge '14},
  location = {Foster City, CA, USA},
  year      = {2014},
  publisher	= {ACM},
  address	= {New York, NY, USA}
}

About

Preference/Ranking Aggregation Library (PAL)

Resources

License

Releases

No releases published

Packages

No packages published

Languages