[mlpack-git] [mlpack/mlpack] Modeling LSH For Performance Tuning (#749)

[Yannis Mentekidis]

This is in no way ready to merge, but I've done a bit of progress on the LSH tuning class, so I'm opening a PR to track how things are going.

With regards to implementation (in order for the algorithm to be complete), I need to

* Write the code that fits Gamma distributions to distances.
* Define the objective functions used for integration to calculate recall and precision.
* Write the code that integrates the objective functions.
* Implement the binary search that identifies useful parameter sets.

With regards to functionality (in order for this to be usable), I also need to

* Identify any hard-coded things that should become input parameters.
* Handle input and output.
* Write meaningful tests.

Then finally there's the step of profiling & optimization of the code.
You can view, comment on, or merge this pull request online at:

  https://github.com/mlpack/mlpack/pull/749

-- Commit Summary --

  * Skeleton of Project
  * Almost complete LSHModel skeleton
  * Implements modeling of arithmetic and geometric mean of distances
  * Merge latest upstream commits into lsh-tuning

-- File Changes --

    M src/mlpack/methods/CMakeLists.txt (1)
    M src/mlpack/methods/lsh/lsh_search.hpp (1)
    A src/mlpack/methods/lsh_model/CMakeLists.txt (19)
    A src/mlpack/methods/lsh_model/lshmodel.hpp (305)
    A src/mlpack/methods/lsh_model/lshmodel_impl.hpp (249)
    A src/mlpack/methods/lsh_model/lshmodel_main.cpp (25)
    A src/mlpack/methods/lsh_model/objectivefunction.hpp (171)

-- Patch Links --

https://github.com/mlpack/mlpack/pull/749.patch
https://github.com/mlpack/mlpack/pull/749.diff

---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/pull/749
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160804/cb1f134a/attachment-0001.html>