<p>In <a href="https://github.com/mlpack/mlpack/pull/746#discussion_r75229770">src/mlpack/core/tree/binary_space_tree/ub_tree_split_impl.hpp</a>:</p>
<pre style='color:#555'>> + const size_t begin,
> + const size_t count,
> + size_t& splitCol)
> +{
> + constexpr size_t order = sizeof(AddressElemType) * CHAR_BIT;
> + if (begin == 0 && count == data.n_cols)
> + {
> + // Calculate all addresses.
> + InitializeAddresses(data);
> +
> + // Probably this is not a good idea. Maybe it is better to get
> + // a number of distinct samples and find the median.
> + std::sort(addresses.begin(), addresses.end(), ComparePair);
> +
> + // Rearrange dataset.
> + PerformSplit(data, count);
</pre>
<p>I think that the code could be simplified significantly if you instead used <code>arma::Mat<AddressElemType></code> in order to store the addresses, and then used <code>sort_index()</code> to get the sorted order, which you could then apply to the data points and the addresses, and use to calculate <code>oldFromNew</code> if needed. I think that would remove the need for <code>PerformSplit</code> altogether.</p>
<p style="font-size:small;-webkit-text-size-adjust:none;color:#666;">—<br />You are receiving this because you are subscribed to this thread.<br />Reply to this email directly, <a href="https://github.com/mlpack/mlpack/pull/746/files/8c5a97dcb1641ae6c98edc70426fb19f5cd7cb79#r75229770">view it on GitHub</a>, or <a href="https://github.com/notifications/unsubscribe-auth/AJ4bFJ_Mg9e7yxiUgmbG49AnWb9OKLxTks5qg6GdgaJpZM4JZrEi">mute the thread</a>.<img alt="" height="1" src="https://github.com/notifications/beacon/AJ4bFOEtMSikrtMdmrt7iFRGQ3bqKhlgks5qg6GdgaJpZM4JZrEi.gif" width="1" /></p>
<div itemscope itemtype="http://schema.org/EmailMessage">
<div itemprop="action" itemscope itemtype="http://schema.org/ViewAction">
<link itemprop="url" href="https://github.com/mlpack/mlpack/pull/746/files/8c5a97dcb1641ae6c98edc70426fb19f5cd7cb79#r75229770"></link>
<meta itemprop="name" content="View Pull Request"></meta>
</div>
<meta itemprop="description" content="View this Pull Request on GitHub"></meta>
</div>
<script type="application/json" data-scope="inboxmarkup">{"api_version":"1.0","publisher":{"api_key":"05dde50f1d1a384dd78767c55493e4bb","name":"GitHub"},"entity":{"external_key":"github/mlpack/mlpack","title":"mlpack/mlpack","subtitle":"GitHub repository","main_image_url":"https://cloud.githubusercontent.com/assets/143418/17495839/a5054eac-5d88-11e6-95fc-7290892c7bb5.png","avatar_image_url":"https://cloud.githubusercontent.com/assets/143418/15842166/7c72db34-2c0b-11e6-9aed-b52498112777.png","action":{"name":"Open in GitHub","url":"https://github.com/mlpack/mlpack"}},"updates":{"snippets":[{"icon":"PERSON","message":"@rcurtin in #746: I think that the code could be simplified significantly if you instead used `arma::Mat\u003cAddressElemType\u003e` in order to store the addresses, and then used `sort_index()` to get the sorted order, which you could then apply to the data points and the addresses, and use to calculate `oldFromNew` if needed. I think that would remove the need for `PerformSplit` altogether."}],"action":{"name":"View Pull Request","url":"https://github.com/mlpack/mlpack/pull/746/files/8c5a97dcb1641ae6c98edc70426fb19f5cd7cb79#r75229770"}}}</script>