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

[Ryan Curtin]

> +    *      k-nearest neighbor. One column per k.
> +    * @param Gk The geometric mean of the squared distances of a point and its
> +    *      k-nearest neighbor. One column per k.
> +    */
> +   void ApproximateKNNStatistics(const arma::Col<size_t>& referenceSizes, 
> +                                 const arma::Col<size_t>& kValues,
> +                                 const arma::mat& Ek, 
> +                                 const arma::mat& Gk);
> +
> +   /**
> +    * This is a helper class that uses the function a * k^b * N^c for some
> +    * parameters a, b, c that have been fit to either predict the arithmetic or
> +    * geometric mean of the squared distance of a point to its k-nearest
> +    * neighbor, given some dataset size N and its k-nearest neighbor.
> +    */
> +   class DistanceStatisticPredictor

If you like, you can make this file smaller by moving `DistanceStatisticPredictor` into a different file and making it not a nested class of `LSHModel`.  I'm ambivalent either way.

-- 
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/files/cdcb575826bfb3bd0ef4cafacf465435b3d6d144#r74688529
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160813/684682f4/attachment.html>