<p>Hi there,</p>

<p>If you are using a discrete distribution (<code>--type discrete</code>), then multidimensional observations aren't supported.  It would not be impossible to support multidimensional discrete distributions (but there will be no relation between dimensions), but currently the code doesn't do that.  I think that we should leave this issue open until multidimensional support is added for HMMs.  For now, you might consider "collapsing" your dimensions: suppose that each of your three dimensions takes values between 0 and 7 (so, 3 bits).  Then given <code>(a, b, c)</code> you map to a single-dimension value of <code>2^6 * a + 2^3 * b + c</code>.  That will cost performance but it will be a workaround for now.  You could also use a Gaussian HMM if the dimensions are not categorical (<code>--type gaussian</code>).</p>

<p>Sorry that the support you are looking for is currently not there.</p>

<p style="font-size:small;-webkit-text-size-adjust:none;color:#666;">&mdash;<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/issues/809#issuecomment-258858129">view it on GitHub</a>, or <a href="https://github.com/notifications/unsubscribe-auth/AJ4bFAhOWxd4xKcw7IVJ_XUL7wTnjCHNks5q7zzjgaJpZM4KpOW0">mute the thread</a>.<img alt="" height="1" src="https://github.com/notifications/beacon/AJ4bFI6nXvbGtF-Cm6fO-rWvkcpr8a4_ks5q7zzjgaJpZM4KpOW0.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/issues/809#issuecomment-258858129"></link>
  <meta itemprop="name" content="View Issue"></meta>
</div>
<meta itemprop="description" content="View this Issue 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 #809: Hi there,\r\n\r\nIf you are using a discrete distribution (`--type discrete`), then multidimensional observations aren't supported.  It would not be impossible to support multidimensional discrete distributions (but there will be no relation between dimensions), but currently the code doesn't do that.  I think that we should leave this issue open until multidimensional support is added for HMMs.  For now, you might consider \"collapsing\" your dimensions: suppose that each of your three dimensions takes values between 0 and 7 (so, 3 bits).  Then given `(a, b, c)` you map to a single-dimension value of `2^6 * a + 2^3 * b + c`.  That will cost performance but it will be a workaround for now.  You could also use a Gaussian HMM if the dimensions are not categorical (`--type gaussian`).\r\n\r\nSorry that the support you are looking for is currently not there."}],"action":{"name":"View Issue","url":"https://github.com/mlpack/mlpack/issues/809#issuecomment-258858129"}}}</script>