{
  "id": 116323,
  "title": "Why is the meter=2 type of measurement so clearly bounded in the training set?",
  "url": "/competitions/ashrae-energy-prediction/discussion/116323",
  "author_name": "",
  "post_date": "2019-11-08T12:10:08.108729900Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3813808%2F45a9b3d71579f3d3f554b108199e5067%2F_20191108200625.png?generation=1573214848243461&amp;alt=media\" alt=\"\">\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3813808%2Fa41c38238aa53f0098af7bfbebd938ad%2F1.png?generation=1573214821573701&amp;alt=media\" alt=\"\"></p>\n\n<p>as we can see, After June 16, the data showed a precipitous drop, which should be abnormal.Why?</p>",
  "messages": [
    {
      "id": "668418",
      "postDate": "11/08/2019 12:10:08",
      "content": "<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3813808%2F45a9b3d71579f3d3f554b108199e5067%2F_20191108200625.png?generation=1573214848243461&amp;alt=media\" alt=\"\">\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3813808%2Fa41c38238aa53f0098af7bfbebd938ad%2F1.png?generation=1573214821573701&amp;alt=media\" alt=\"\"></p>\n\n<p>as we can see, After June 16, the data showed a precipitous drop, which should be abnormal.Why?</p>",
      "rawMarkdown": "![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3813808%2F45a9b3d71579f3d3f554b108199e5067%2F_20191108200625.png?generation=1573214848243461&amp;alt=media)\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3813808%2Fa41c38238aa53f0098af7bfbebd938ad%2F1.png?generation=1573214821573701&amp;alt=media)\n\nas we can see, After June 16, the data showed a precipitous drop, which should be abnormal.Why?",
      "votes": null
    },
    {
      "id": "668422",
      "postDate": "11/08/2019 12:15:54",
      "content": "<p>Missing values</p>",
      "rawMarkdown": "Missing values",
      "votes": null
    },
    {
      "id": "668880",
      "postDate": "11/09/2019 02:09:12",
      "content": "<p>thank you ,i think it may caused by building1099...And when I was training, rmse could reach 0.6, which is very serious in over-fitting. Is this related to my use of many new features to train?Based on other people's models, it doesn't seem like you need a lot of features to train the model</p>",
      "rawMarkdown": "thank you ,i think it may caused by building1099...And when I was training, rmse could reach 0.6, which is very serious in over-fitting. Is this related to my use of many new features to train?Based on other people's models, it doesn't seem like you need a lot of features to train the model",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 668422,
      "author_name": "jielu0728",
      "author_url": "",
      "post_date": "11/08/2019 12:15:54",
      "content": "<p>Missing values</p>",
      "votes": null,
      "replies": [
        {
          "id": 668880,
          "author_name": "xhzs2333",
          "author_url": "",
          "post_date": "11/09/2019 02:09:12",
          "content": "<p>thank you ,i think it may caused by building1099...And when I was training, rmse could reach 0.6, which is very serious in over-fitting. Is this related to my use of many new features to train?Based on other people's models, it doesn't seem like you need a lot of features to train the model</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "668418": "![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3813808%2F45a9b3d71579f3d3f554b108199e5067%2F_20191108200625.png?generation=1573214848243461&amp;alt=media)\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3813808%2Fa41c38238aa53f0098af7bfbebd938ad%2F1.png?generation=1573214821573701&amp;alt=media)\n\nas we can see, After June 16, the data showed a precipitous drop, which should be abnormal.Why?",
    "668422": "Missing values",
    "668880": "thank you ,i think it may caused by building1099...And when I was training, rmse could reach 0.6, which is very serious in over-fitting. Is this related to my use of many new features to train?Based on other people's models, it doesn't seem like you need a lot of features to train the model"
  },
  "source": "meta"
}