{
  "id": 126795,
  "title": "False positives not included?",
  "url": "/competitions/pku-autonomous-driving/discussion/126795",
  "author_name": "",
  "post_date": "2020-01-20T10:06:32.747405800Z",
  "votes": 4,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi guys, last two days, wish you all luck.\nI suspect that the false positive did't include in metric estimation in current LB.\nDoes anyone feel the same?</p>",
  "messages": [
    {
      "id": "723673",
      "postDate": "01/20/2020 10:06:32",
      "content": "<p>Hi guys, last two days, wish you all luck.\nI suspect that the false positive did't include in metric estimation in current LB.\nDoes anyone feel the same?</p>",
      "rawMarkdown": "Hi guys, last two days, wish you all luck.\nI suspect that the false positive did't include in metric estimation in current LB.\nDoes anyone feel the same?",
      "votes": null
    },
    {
      "id": "723808",
      "postDate": "01/20/2020 13:37:21",
      "content": "<p>There are other problems too and admins never told us what they are actually doing...</p>",
      "rawMarkdown": "There are other problems too and admins never told us what they are actually doing...",
      "votes": null
    },
    {
      "id": "723817",
      "postDate": "01/20/2020 13:48:21",
      "content": "<p>I have noticed this since I visualize the information in the submission csv. And I found there are several cases that duplicate predictions on the same car. But the score still much better than the submission file which doesn't have this problem. Although this weird submission file is coming from ensemble two models, but I think the score shouldn't be better if they include false positive in it...</p>",
      "rawMarkdown": "I have noticed this since I visualize the information in the submission csv. And I found there are several cases that duplicate predictions on the same car. But the score still much better than the submission file which doesn't have this problem. Although this weird submission file is coming from ensemble two models, but I think the score shouldn't be better if they include false positive in it...",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 723808,
      "author_name": "greatgamedota",
      "author_url": "",
      "post_date": "01/20/2020 13:37:21",
      "content": "<p>There are other problems too and admins never told us what they are actually doing...</p>",
      "votes": null,
      "replies": [
        {
          "id": 723817,
          "author_name": "xiejialun",
          "author_url": "",
          "post_date": "01/20/2020 13:48:21",
          "content": "<p>I have noticed this since I visualize the information in the submission csv. And I found there are several cases that duplicate predictions on the same car. But the score still much better than the submission file which doesn't have this problem. Although this weird submission file is coming from ensemble two models, but I think the score shouldn't be better if they include false positive in it...</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "723673": "Hi guys, last two days, wish you all luck.\nI suspect that the false positive did't include in metric estimation in current LB.\nDoes anyone feel the same?",
    "723808": "There are other problems too and admins never told us what they are actually doing...",
    "723817": "I have noticed this since I visualize the information in the submission csv. And I found there are several cases that duplicate predictions on the same car. But the score still much better than the submission file which doesn't have this problem. Although this weird submission file is coming from ensemble two models, but I think the score shouldn't be better if they include false positive in it..."
  },
  "source": "meta"
}