{
  "id": 347503,
  "title": "Hard to push the score  >.78 . Curious to know the trick to uplift it further?",
  "url": "/competitions/amex-default-prediction/discussion/347503",
  "author_name": "",
  "post_date": "2022-08-24T12:21:05.136234300Z",
  "votes": 2,
  "comment_count": 5,
  "views": 0,
  "content": "<p>I have been trying to uplift the performance from &gt;0.78 but none of the method have worked well to push it higher. Since the competition is about to finish now. Would someone mind to explain what was the key thing which uplifted it further. Specifically with the leader board kagglers?</p>",
  "messages": [
    {
      "id": "1911975",
      "postDate": "08/24/2022 12:21:05",
      "content": "<p>I have been trying to uplift the performance from &gt;0.78 but none of the method have worked well to push it higher. Since the competition is about to finish now. Would someone mind to explain what was the key thing which uplifted it further. Specifically with the leader board kagglers?</p>",
      "rawMarkdown": "I have been trying to uplift the performance from >0.78 but none of the method have worked well to push it higher. Since the competition is about to finish now. Would someone mind to explain what was the key thing which uplifted it further. Specifically with the leader board kagglers?",
      "votes": null
    },
    {
      "id": "1911985",
      "postDate": "08/24/2022 12:33:04",
      "content": "<p>There were public notebooks with a score of 0.795 last time I checked, a while ago. I would not be surprised if there were even better ones now.</p>",
      "rawMarkdown": "There were public notebooks with a score of 0.795 last time I checked, a while ago. I would not be surprised if there were even better ones now.",
      "votes": null
    },
    {
      "id": "1911986",
      "postDate": "08/24/2022 12:34:47",
      "content": "<p>We will all find our tomorrow when it finishes.</p>",
      "rawMarkdown": "We will all find our tomorrow when it finishes.",
      "votes": null
    },
    {
      "id": "1912058",
      "postDate": "08/24/2022 13:32:18",
      "content": "<p>I am a new bee in kaggle. But I spent much time to think what's the common thing about GM.</p>\n<p>Yes.</p>\n<p>It </p>\n<p>is </p>\n<p>excellent feature engineering   :D </p>",
      "rawMarkdown": "I am a new bee in kaggle. But I spent much time to think what's the common thing about GM.\n\nYes.\n\nIt \n\nis \n\nexcellent feature engineering   :D",
      "votes": null
    },
    {
      "id": "1912164",
      "postDate": "08/24/2022 14:52:18",
      "content": "<p>model ensembling</p>",
      "rawMarkdown": "model ensembling",
      "votes": null
    },
    {
      "id": "1912690",
      "postDate": "08/24/2022 23:49:44",
      "content": "<p>I think the greatest method in this competition was feature engineering around client's last records. it's all about time series and models blendings in the end</p>",
      "rawMarkdown": "I think the greatest method in this competition was feature engineering around client's last records. it's all about time series and models blendings in the end",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1911985,
      "author_name": "cpmpml",
      "author_url": "",
      "post_date": "08/24/2022 12:33:04",
      "content": "<p>There were public notebooks with a score of 0.795 last time I checked, a while ago. I would not be surprised if there were even better ones now.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1911986,
      "author_name": "jakelj",
      "author_url": "",
      "post_date": "08/24/2022 12:34:47",
      "content": "<p>We will all find our tomorrow when it finishes.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1912058,
      "author_name": "zhehaoliang",
      "author_url": "",
      "post_date": "08/24/2022 13:32:18",
      "content": "<p>I am a new bee in kaggle. But I spent much time to think what's the common thing about GM.</p>\n<p>Yes.</p>\n<p>It </p>\n<p>is </p>\n<p>excellent feature engineering   :D </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1912164,
      "author_name": "nangaparbat",
      "author_url": "",
      "post_date": "08/24/2022 14:52:18",
      "content": "<p>model ensembling</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1912690,
      "author_name": "leandrodestefani",
      "author_url": "",
      "post_date": "08/24/2022 23:49:44",
      "content": "<p>I think the greatest method in this competition was feature engineering around client's last records. it's all about time series and models blendings in the end</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1911975": "I have been trying to uplift the performance from >0.78 but none of the method have worked well to push it higher. Since the competition is about to finish now. Would someone mind to explain what was the key thing which uplifted it further. Specifically with the leader board kagglers?",
    "1911985": "There were public notebooks with a score of 0.795 last time I checked, a while ago. I would not be surprised if there were even better ones now.",
    "1911986": "We will all find our tomorrow when it finishes.",
    "1912058": "I am a new bee in kaggle. But I spent much time to think what's the common thing about GM.\n\nYes.\n\nIt \n\nis \n\nexcellent feature engineering   :D",
    "1912164": "model ensembling",
    "1912690": "I think the greatest method in this competition was feature engineering around client's last records. it's all about time series and models blendings in the end"
  },
  "source": "meta"
}