{
  "id": 74404,
  "title": "Mislabeled?",
  "url": "/competitions/PLAsTiCC-2018/discussion/74404",
  "author_name": "",
  "post_date": "2018-12-12T00:43:37.883981100Z",
  "votes": 3,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Check out these oids: <code>[ 28915,  84306, 161432, 168989, 188157, 200949, 223940, 243320, 281669, 331174 ]</code></p>\n\n<p>Supposedly they are AGN / Class=16 but they don't look like AGNs to me. They look more like variable stars. I haven't attempted to calculate a period for them yet to prove it though. Perhaps relabeling these might bolster score?</p>",
  "messages": [
    {
      "id": "437449",
      "postDate": "12/12/2018 00:43:37",
      "content": "<p>Check out these oids: <code>[ 28915,  84306, 161432, 168989, 188157, 200949, 223940, 243320, 281669, 331174 ]</code></p>\n\n<p>Supposedly they are AGN / Class=16 but they don't look like AGNs to me. They look more like variable stars. I haven't attempted to calculate a period for them yet to prove it though. Perhaps relabeling these might bolster score?</p>",
      "rawMarkdown": "Check out these oids: `[ 28915,  84306, 161432, 168989, 188157, 200949, 223940, 243320, 281669, 331174 ]`\n\nSupposedly they are AGN / Class=16 but they don't look like AGNs to me. They look more like variable stars. I haven't attempted to calculate a period for them yet to prove it though. Perhaps relabeling these might bolster score?",
      "votes": null
    },
    {
      "id": "437464",
      "postDate": "12/12/2018 01:23:12",
      "content": "<p>I don't have any domain knowledge on this but since the data is simulated, I would assume there shouldn't be any mislabelling? it could be small probability events when drawing samples from the ground truth distribution.</p>",
      "rawMarkdown": "I don't have any domain knowledge on this but since the data is simulated, I would assume there shouldn't be any mislabelling? it could be small probability events when drawing samples from the ground truth distribution.",
      "votes": null
    },
    {
      "id": "437472",
      "postDate": "12/12/2018 01:35:08",
      "content": "<p>I thought so too, but the differences are just too pronounced:</p>\n\n<p><img src=\"https://i.imgur.com/rKd6yHL.png\" alt=\"exa\"></p>\n\n<p>VS</p>\n\n<p><img src=\"https://i.imgur.com/3wavSbY.png\" alt=\"exb\"></p>",
      "rawMarkdown": "I thought so too, but the differences are just too pronounced:\n\n![exa](https://i.imgur.com/rKd6yHL.png)\n\nVS\n\n![exb](https://i.imgur.com/3wavSbY.png)",
      "votes": null
    },
    {
      "id": "437524",
      "postDate": "12/12/2018 04:46:36",
      "content": "<p>I really doubt class 16 is AGN given it is a galactic source ...  AGNs are extra galactic by definition given they are a whole galaxy center being seen as a source.  <a href=\"https://en.wikipedia.org/wiki/Active_galactic_nucleus\">https://en.wikipedia.org/wiki/Active_galactic_nucleus</a></p>\n\n<p>I'd wait for Kyle Boone (or any other astronomer here) confirmation, but I'm pretty confident.  </p>",
      "rawMarkdown": "I really doubt class 16 is AGN given it is a galactic source ...  AGNs are extra galactic by definition given they are a whole galaxy center being seen as a source.  https://en.wikipedia.org/wiki/Active_galactic_nucleus\n\nI'd wait for Kyle Boone (or any other astronomer here) confirmation, but I'm pretty confident.",
      "votes": null
    },
    {
      "id": "437558",
      "postDate": "12/12/2018 05:58:29",
      "content": "<p>Looks like I mislabeled internally on my spreadsheet. I think it's some kind of variable star, potentially with planets or other occluding bodies. </p>",
      "rawMarkdown": "Looks like I mislabeled internally on my spreadsheet. I think it's some kind of variable star, potentially with planets or other occluding bodies.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 437464,
      "author_name": "jiweiliu",
      "author_url": "",
      "post_date": "12/12/2018 01:23:12",
      "content": "<p>I don't have any domain knowledge on this but since the data is simulated, I would assume there shouldn't be any mislabelling? it could be small probability events when drawing samples from the ground truth distribution.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 437472,
      "author_name": "authman",
      "author_url": "",
      "post_date": "12/12/2018 01:35:08",
      "content": "<p>I thought so too, but the differences are just too pronounced:</p>\n\n<p><img src=\"https://i.imgur.com/rKd6yHL.png\" alt=\"exa\"></p>\n\n<p>VS</p>\n\n<p><img src=\"https://i.imgur.com/3wavSbY.png\" alt=\"exb\"></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 437524,
      "author_name": "cpmpml",
      "author_url": "",
      "post_date": "12/12/2018 04:46:36",
      "content": "<p>I really doubt class 16 is AGN given it is a galactic source ...  AGNs are extra galactic by definition given they are a whole galaxy center being seen as a source.  <a href=\"https://en.wikipedia.org/wiki/Active_galactic_nucleus\">https://en.wikipedia.org/wiki/Active_galactic_nucleus</a></p>\n\n<p>I'd wait for Kyle Boone (or any other astronomer here) confirmation, but I'm pretty confident.  </p>",
      "votes": null,
      "replies": [
        {
          "id": 437558,
          "author_name": "authman",
          "author_url": "",
          "post_date": "12/12/2018 05:58:29",
          "content": "<p>Looks like I mislabeled internally on my spreadsheet. I think it's some kind of variable star, potentially with planets or other occluding bodies. </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "437449": "Check out these oids: `[ 28915,  84306, 161432, 168989, 188157, 200949, 223940, 243320, 281669, 331174 ]`\n\nSupposedly they are AGN / Class=16 but they don't look like AGNs to me. They look more like variable stars. I haven't attempted to calculate a period for them yet to prove it though. Perhaps relabeling these might bolster score?",
    "437464": "I don't have any domain knowledge on this but since the data is simulated, I would assume there shouldn't be any mislabelling? it could be small probability events when drawing samples from the ground truth distribution.",
    "437472": "I thought so too, but the differences are just too pronounced:\n\n![exa](https://i.imgur.com/rKd6yHL.png)\n\nVS\n\n![exb](https://i.imgur.com/3wavSbY.png)",
    "437524": "I really doubt class 16 is AGN given it is a galactic source ...  AGNs are extra galactic by definition given they are a whole galaxy center being seen as a source.  https://en.wikipedia.org/wiki/Active_galactic_nucleus\n\nI'd wait for Kyle Boone (or any other astronomer here) confirmation, but I'm pretty confident.",
    "437558": "Looks like I mislabeled internally on my spreadsheet. I think it's some kind of variable star, potentially with planets or other occluding bodies."
  },
  "source": "meta"
}