{
  "id": 47708,
  "title": " what do you think about possible performance we can achieve after three months?",
  "url": "/competitions/data-science-bowl-2018/discussion/47708",
  "author_name": "",
  "post_date": "2018-01-17T21:56:06.315768400Z",
  "votes": 4,
  "comment_count": 9,
  "views": 0,
  "content": "<p>Good night, my friends. At the beginning of competition.... what do you think about possible performance we can achieve after three months?</p>",
  "messages": [
    {
      "id": "270183",
      "postDate": "01/17/2018 21:56:06",
      "content": "<p>Good night, my friends. At the beginning of competition.... what do you think about possible performance we can achieve after three months?</p>",
      "rawMarkdown": "Good night, my friends. At the beginning of competition.... what do you think about possible performance we can achieve after three months?",
      "votes": null
    },
    {
      "id": "270185",
      "postDate": "01/17/2018 22:01:10",
      "content": "<p>make your bets )</p>",
      "rawMarkdown": "make your bets )",
      "votes": null
    },
    {
      "id": "270240",
      "postDate": "01/18/2018 00:13:02",
      "content": "<p>I think the private leaderboard will peak at 0.85. As for the public lb.... we’ll have to see how much overfitting is done</p>",
      "rawMarkdown": "I think the private leaderboard will peak at 0.85. As for the public lb.... we’ll have to see how much overfitting is done",
      "votes": null
    },
    {
      "id": "270494",
      "postDate": "01/18/2018 10:45:25",
      "content": "<p>Private and public LBs are same in this competition.</p>",
      "rawMarkdown": "Private and public LBs are same in this competition.",
      "votes": null
    },
    {
      "id": "270560",
      "postDate": "01/18/2018 14:19:36",
      "content": "<p>I really don’t know. SOA for natural images using the same metric is 0.46. It’s a simpler problem (e.g. less classes and plainer images) but robustness is more difficult. I’d guess 0.7 (but I’m an optimist).</p>",
      "rawMarkdown": "I really don’t know. SOA for natural images using the same metric is 0.46. It’s a simpler problem (e.g. less classes and plainer images) but robustness is more difficult. I’d guess 0.7 (but I’m an optimist).",
      "votes": null
    },
    {
      "id": "270582",
      "postDate": "01/18/2018 14:55:24",
      "content": "<p>@Vivek this is a two-stage competition, so the private leaderboard will not be the same once the second stage occurs. See - <a href=\"https://www.kaggle.com/two-stage-faq\">https://www.kaggle.com/two-stage-faq</a></p>",
      "rawMarkdown": "Vivek this is a two-stage competition, so the private leaderboard will not be the same once the second stage occurs. See - https://www.kaggle.com/two-stage-faq",
      "votes": null
    },
    {
      "id": "270626",
      "postDate": "01/18/2018 16:20:28",
      "content": "<p>Aah, thanks!</p>",
      "rawMarkdown": "Aah, thanks!",
      "votes": null
    },
    {
      "id": "270673",
      "postDate": "01/18/2018 18:03:59",
      "content": "<p>Sorry, William. What do you think about final performance?</p>",
      "rawMarkdown": "Sorry, William. What do you think about final performance?",
      "votes": null
    },
    {
      "id": "270802",
      "postDate": "01/19/2018 00:59:46",
      "content": "<p>What's SOA?</p>",
      "rawMarkdown": "What's SOA?",
      "votes": null
    },
    {
      "id": "271140",
      "postDate": "01/19/2018 19:12:45",
      "content": "<p>state of the art ?</p>",
      "rawMarkdown": "state of the art ?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 270185,
      "author_name": "alexanderkireev",
      "author_url": "",
      "post_date": "01/17/2018 22:01:10",
      "content": "<p>make your bets )</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 270240,
      "author_name": "jzerox2",
      "author_url": "",
      "post_date": "01/18/2018 00:13:02",
      "content": "<p>I think the private leaderboard will peak at 0.85. As for the public lb.... we’ll have to see how much overfitting is done</p>",
      "votes": null,
      "replies": [
        {
          "id": 270494,
          "author_name": "thevivekpandey",
          "author_url": "",
          "post_date": "01/18/2018 10:45:25",
          "content": "<p>Private and public LBs are same in this competition.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 270582,
          "author_name": "wcukierski",
          "author_url": "",
          "post_date": "01/18/2018 14:55:24",
          "content": "<p>@Vivek this is a two-stage competition, so the private leaderboard will not be the same once the second stage occurs. See - <a href=\"https://www.kaggle.com/two-stage-faq\">https://www.kaggle.com/two-stage-faq</a></p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 270626,
          "author_name": "thevivekpandey",
          "author_url": "",
          "post_date": "01/18/2018 16:20:28",
          "content": "<p>Aah, thanks!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 270673,
          "author_name": "alexanderkireev",
          "author_url": "",
          "post_date": "01/18/2018 18:03:59",
          "content": "<p>Sorry, William. What do you think about final performance?</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 270560,
      "author_name": "agoodman",
      "author_url": "",
      "post_date": "01/18/2018 14:19:36",
      "content": "<p>I really don’t know. SOA for natural images using the same metric is 0.46. It’s a simpler problem (e.g. less classes and plainer images) but robustness is more difficult. I’d guess 0.7 (but I’m an optimist).</p>",
      "votes": null,
      "replies": [
        {
          "id": 270802,
          "author_name": "pauljurczak",
          "author_url": "",
          "post_date": "01/19/2018 00:59:46",
          "content": "<p>What's SOA?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 271140,
          "author_name": "witeko",
          "author_url": "",
          "post_date": "01/19/2018 19:12:45",
          "content": "<p>state of the art ?</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "270183": "Good night, my friends. At the beginning of competition.... what do you think about possible performance we can achieve after three months?",
    "270185": "make your bets )",
    "270240": "I think the private leaderboard will peak at 0.85. As for the public lb.... we’ll have to see how much overfitting is done",
    "270494": "Private and public LBs are same in this competition.",
    "270560": "I really don’t know. SOA for natural images using the same metric is 0.46. It’s a simpler problem (e.g. less classes and plainer images) but robustness is more difficult. I’d guess 0.7 (but I’m an optimist).",
    "270582": "Vivek this is a two-stage competition, so the private leaderboard will not be the same once the second stage occurs. See - https://www.kaggle.com/two-stage-faq",
    "270626": "Aah, thanks!",
    "270673": "Sorry, William. What do you think about final performance?",
    "270802": "What's SOA?",
    "271140": "state of the art ?"
  },
  "source": "meta"
}