{
  "id": 76320,
  "title": "Is it important to distinguish outliers?",
  "url": "/competitions/elo-merchant-category-recommendation/discussion/76320",
  "author_name": "",
  "post_date": "2019-01-01T08:47:47.316139300Z",
  "votes": null,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Is it important to distinguish outliers?\nIf important, I will ask for a model to distinguish outliers.</p>",
  "messages": [
    {
      "id": "448465",
      "postDate": "01/01/2019 08:47:47",
      "content": "<p>Is it important to distinguish outliers?\nIf important, I will ask for a model to distinguish outliers.</p>",
      "rawMarkdown": "Is it important to distinguish outliers?\nIf important, I will ask for a model to distinguish outliers.",
      "votes": null
    },
    {
      "id": "448617",
      "postDate": "01/01/2019 17:08:59",
      "content": "<p>Yes. <a href=\"https://www.kaggle.com/waitingli/combining-your-model-with-a-model-without-outlier\">See here for an example.</a></p>",
      "rawMarkdown": "Yes. [See here for an example.](https://www.kaggle.com/waitingli/combining-your-model-with-a-model-without-outlier)",
      "votes": null
    },
    {
      "id": "448779",
      "postDate": "01/02/2019 04:42:09",
      "content": "<p>I read that kernel.\nHowever, the prediction accuracy of outliers does not appear to be that good. So I am still looking for a model with good classification accuracy.</p>",
      "rawMarkdown": "I read that kernel.\nHowever, the prediction accuracy of outliers does not appear to be that good. So I am still looking for a model with good classification accuracy.",
      "votes": null
    },
    {
      "id": "448822",
      "postDate": "01/02/2019 07:02:52",
      "content": "<p>Guess you may have to make one yourself! ;)</p>",
      "rawMarkdown": "Guess you may have to make one yourself! ;)",
      "votes": null
    },
    {
      "id": "448841",
      "postDate": "01/02/2019 07:58:59",
      "content": "<p>I agree. I will make a model and try to compare the correct answer rate with the previous kernel.</p>",
      "rawMarkdown": "I agree. I will make a model and try to compare the correct answer rate with the previous kernel.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 448617,
      "author_name": "peterhurford",
      "author_url": "",
      "post_date": "01/01/2019 17:08:59",
      "content": "<p>Yes. <a href=\"https://www.kaggle.com/waitingli/combining-your-model-with-a-model-without-outlier\">See here for an example.</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 448779,
      "author_name": "",
      "author_url": "",
      "post_date": "01/02/2019 04:42:09",
      "content": "<p>I read that kernel.\nHowever, the prediction accuracy of outliers does not appear to be that good. So I am still looking for a model with good classification accuracy.</p>",
      "votes": null,
      "replies": [
        {
          "id": 448822,
          "author_name": "peterhurford",
          "author_url": "",
          "post_date": "01/02/2019 07:02:52",
          "content": "<p>Guess you may have to make one yourself! ;)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 448841,
          "author_name": "",
          "author_url": "",
          "post_date": "01/02/2019 07:58:59",
          "content": "<p>I agree. I will make a model and try to compare the correct answer rate with the previous kernel.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "448465": "Is it important to distinguish outliers?\nIf important, I will ask for a model to distinguish outliers.",
    "448617": "Yes. [See here for an example.](https://www.kaggle.com/waitingli/combining-your-model-with-a-model-without-outlier)",
    "448779": "I read that kernel.\nHowever, the prediction accuracy of outliers does not appear to be that good. So I am still looking for a model with good classification accuracy.",
    "448822": "Guess you may have to make one yourself! ;)",
    "448841": "I agree. I will make a model and try to compare the correct answer rate with the previous kernel."
  },
  "source": "meta"
}