Bitbucket is a code hosting site with unlimited public and private repositories. We're also free for small teams!

This package contains the scripts and sources to drive Michael
Ekstrand and John Riedl's RecSys 2012 paper “When Recommenders Fail:
Identifying and Predicting Recommender Failure for Evidence-Based
Algorithm Selection and Combination”.


- LensKit 0.11-SNAPSHOT, from,
  revision f0215b1985f2 with the lenskit-clamp-predictors.patch patch
  (in this package) applied.

- Java 6 and Maven 3

- R (tested with 2.14.2; requires a version with the 'parallel' package)

- R packages
  - doBy
  - xtable
  - ROCR
  - ggplot2

- Emacs with Org-Mode 7.8.0 (to re-export the R scripts)

To run:

1. Clone LensKit
   hg clone -r f0215b1985f2

2. Apply the patch (in LensKit dir)
   patch -p1 <../error-analysis/lenskit-clamp-predictors.patch

3. Install LensKit (in LensKit dir)
   mvn install

4. Download the ML-10M data set from and
   unpack it into a directory called ‘ml-10m’ (the 'ratings.dat' file
   should reside directly in ‘ml-10m’).

5. Compile & build the add-on code
   mvn package

6. Import the MovieLens ratings into a Lucene index
   ./target/dist/bin/import-tags ml-10m/movies.dat ml-10m/tags.dat ml-10m.idx

7. Run the recommenders (-j4 uses 4 threads)
   ./target/dist/bin/lenskit-eval -j4 lenskit.groovy

8. Run the R scripts (runs with library & data load dependencies)
   cd paper && ./ *.R

9. Build the paper in paper/error-analysis-short.tex (requires
   PDFLaTeX and BibTeX)

Recent activity

Tip: Filter by directory path e.g. /media app.js to search for public/media/app.js.
Tip: Use camelCasing e.g. ProjME to search for
Tip: Filter by extension type e.g. /repo .js to search for all .js files in the /repo directory.
Tip: Separate your search with spaces e.g. /ssh pom.xml to search for src/ssh/pom.xml.
Tip: Use ↑ and ↓ arrow keys to navigate and return to view the file.
Tip: You can also navigate files with Ctrl+j (next) and Ctrl+k (previous) and view the file with Ctrl+o.
Tip: You can also navigate files with Alt+j (next) and Alt+k (previous) and view the file with Alt+o.