{
  "id": 17022,
  "title": "Can we see the training set with all available features",
  "url": "/competitions/flavours-of-physics/discussion/17022",
  "author_name": "",
  "post_date": "2015-10-15T11:22:54.747Z",
  "votes": 3,
  "comment_count": 1,
  "views": 267,
  "content": "<p>As we all know a lot of features were removed from data set to make this ML research as useful from physical perspective as possible. But now, after competition, many of us are still doing some research on data. I just thought it would be useful for us all to see all parameters that we can read from detector. I am especially interested in all momentum vectors as I got very puzzling results in my angles approximation approach. So please, organizers, can you provide the training set with all available parameters?</p>",
  "messages": [
    {
      "id": "96262",
      "postDate": "10/15/2015 11:22:54",
      "content": "<p>As we all know a lot of features were removed from data set to make this ML research as useful from physical perspective as possible. But now, after competition, many of us are still doing some research on data. I just thought it would be useful for us all to see all parameters that we can read from detector. I am especially interested in all momentum vectors as I got very puzzling results in my angles approximation approach. So please, organizers, can you provide the training set with all available parameters?</p>",
      "rawMarkdown": "As we all know a lot of features were removed from data set to make this ML research as useful from physical perspective as possible. But now, after competition, many of us are still doing some research on data. I just thought it would be useful for us all to see all parameters that we can read from detector. I am especially interested in all momentum vectors as I got very puzzling results in my angles approximation approach. So please, organizers, can you provide the training set with all available parameters?",
      "votes": null
    },
    {
      "id": "96264",
      "postDate": "10/15/2015 11:33:02",
      "content": "<p>I totally support this request. I'm sure that at least azimuthal angles (phi) were removed from the data set, and this limits our ability to use kinematics to distinguish between signal and background, but doesn't really limit our ability to reconstruct invariant masses. It would be great to make the original data set freely available, for example, through the CERN Open Data website (<a href=\"http://opendata.cern.ch/\">http://opendata.cern.ch</a>).</p>",
      "rawMarkdown": "I totally support this request. I'm sure that at least azimuthal angles (phi) were removed from the data set, and this limits our ability to use kinematics to distinguish between signal and background, but doesn't really limit our ability to reconstruct invariant masses. It would be great to make the original data set freely available, for example, through the CERN Open Data website (http://opendata.cern.ch).",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 96264,
      "author_name": "gramolin",
      "author_url": "",
      "post_date": "10/15/2015 11:33:02",
      "content": "<p>I totally support this request. I'm sure that at least azimuthal angles (phi) were removed from the data set, and this limits our ability to use kinematics to distinguish between signal and background, but doesn't really limit our ability to reconstruct invariant masses. It would be great to make the original data set freely available, for example, through the CERN Open Data website (<a href=\"http://opendata.cern.ch/\">http://opendata.cern.ch</a>).</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "96262": "As we all know a lot of features were removed from data set to make this ML research as useful from physical perspective as possible. But now, after competition, many of us are still doing some research on data. I just thought it would be useful for us all to see all parameters that we can read from detector. I am especially interested in all momentum vectors as I got very puzzling results in my angles approximation approach. So please, organizers, can you provide the training set with all available parameters?",
    "96264": "I totally support this request. I'm sure that at least azimuthal angles (phi) were removed from the data set, and this limits our ability to use kinematics to distinguish between signal and background, but doesn't really limit our ability to reconstruct invariant masses. It would be great to make the original data set freely available, for example, through the CERN Open Data website (http://opendata.cern.ch)."
  },
  "source": "meta"
}