{
  "id": 97163,
  "title": "Addition data not present for Test Data",
  "url": "/competitions/champs-scalar-coupling/discussion/97163",
  "author_name": "",
  "post_date": "2019-06-26T02:06:46.361788500Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>As stated the test data does not contain additional info. So if I train my training data by adding the features from these additional data, how am I going to construct similar features for the test? Do I need the figure out a way to link the columns in test data(except for molecule_name) with the additional data provided for train set?</p>\n\n<p>Please help.\nThanks</p>",
  "messages": [
    {
      "id": "560970",
      "postDate": "06/26/2019 02:06:46",
      "content": "<p>As stated the test data does not contain additional info. So if I train my training data by adding the features from these additional data, how am I going to construct similar features for the test? Do I need the figure out a way to link the columns in test data(except for molecule_name) with the additional data provided for train set?</p>\n\n<p>Please help.\nThanks</p>",
      "rawMarkdown": "As stated the test data does not contain additional info. So if I train my training data by adding the features from these additional data, how am I going to construct similar features for the test? Do I need the figure out a way to link the columns in test data(except for molecule_name) with the additional data provided for train set?\n\nPlease help.\nThanks",
      "votes": null
    },
    {
      "id": "561043",
      "postDate": "06/26/2019 04:38:39",
      "content": "<p>Please refer to this thread : <a href=\"https://www.kaggle.com/c/champs-scalar-coupling/discussion/96605\">https://www.kaggle.com/c/champs-scalar-coupling/discussion/96605</a></p>",
      "rawMarkdown": "Please refer to this thread : https://www.kaggle.com/c/champs-scalar-coupling/discussion/96605",
      "votes": null
    },
    {
      "id": "569420",
      "postDate": "07/06/2019 15:54:02",
      "content": "<p>Did you figure it out? I still don't get it. I would say yes, train different models on these new features so we can generate them for the test set but I'm not sure</p>",
      "rawMarkdown": "Did you figure it out? I still don't get it. I would say yes, train different models on these new features so we can generate them for the test set but I'm not sure",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 561043,
      "author_name": "arnabbiswas1",
      "author_url": "",
      "post_date": "06/26/2019 04:38:39",
      "content": "<p>Please refer to this thread : <a href=\"https://www.kaggle.com/c/champs-scalar-coupling/discussion/96605\">https://www.kaggle.com/c/champs-scalar-coupling/discussion/96605</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 569420,
      "author_name": "lpintoc",
      "author_url": "",
      "post_date": "07/06/2019 15:54:02",
      "content": "<p>Did you figure it out? I still don't get it. I would say yes, train different models on these new features so we can generate them for the test set but I'm not sure</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "560970": "As stated the test data does not contain additional info. So if I train my training data by adding the features from these additional data, how am I going to construct similar features for the test? Do I need the figure out a way to link the columns in test data(except for molecule_name) with the additional data provided for train set?\n\nPlease help.\nThanks",
    "561043": "Please refer to this thread : https://www.kaggle.com/c/champs-scalar-coupling/discussion/96605",
    "569420": "Did you figure it out? I still don't get it. I would say yes, train different models on these new features so we can generate them for the test set but I'm not sure"
  },
  "source": "meta"
}