/LtR.jl

Source code of SIGIR'17 paper "Learning to rank using localized geometry mean metrics"

Primary LanguageJulia

Source code of SIGIR'17 paper "Learning to rank using localized geometry mean metrics"

(ACM version, Arxiv version)

The program is written by Julia. Please download the lastest version to execute this demo code.

Prerequisites

You need to install two packages MAT and Iterators in the command shell of Julia:

julia > Pkg.add("MAT")
julia > Pkg.add("Iterators")

Execution

We have prepared a demo program to train two rank lists from MSLR-WEB10K dataset. You can try our program under the project directory and type:

julia demo_small_dataset.jl

Citation

Please cite this paper if you use the source code with academic purpose:

@inproceedings{DBLP:conf/sigir/SuKL17,
  author    = {Yuxin Su and
               Irwin King and
               Michael R. Lyu},
  title     = {Learning to Rank Using Localized Geometric Mean Metrics},
  booktitle = {Proceedings of the 40th International {ACM} {SIGIR} Conference on
               Research and Development in Information Retrieval, Shinjuku, Tokyo,
               Japan, August 7-11, 2017},
  pages     = {45--54},
  year      = {2017},
  crossref  = {DBLP:conf/sigir/2017},
  url       = {http://doi.acm.org/10.1145/3077136.3080828},
  doi       = {10.1145/3077136.3080828},
  timestamp = {Sun, 06 Aug 2017 18:21:32 +0200},
  biburl    = {http://dblp.org/rec/bib/conf/sigir/SuKL17},
  bibsource = {dblp computer science bibliography, http://dblp.org}
}