{
  "id": 94704,
  "title": "Data real or from simulation?",
  "url": "/competitions/champs-scalar-coupling/discussion/94704",
  "author_name": "",
  "post_date": "2019-06-06T10:28:59.238862200Z",
  "votes": 4,
  "comment_count": 6,
  "views": 0,
  "content": "<p>I just read the description and checked discussion threads. But cannot find the answer. So it this dataset from a chemistry simulation/theory/formulas or are they from experimental results?</p>",
  "messages": [
    {
      "id": "546177",
      "postDate": "06/06/2019 10:28:59",
      "content": "<p>I just read the description and checked discussion threads. But cannot find the answer. So it this dataset from a chemistry simulation/theory/formulas or are they from experimental results?</p>",
      "rawMarkdown": "I just read the description and checked discussion threads. But cannot find the answer. So it this dataset from a chemistry simulation/theory/formulas or are they from experimental results?",
      "votes": null
    },
    {
      "id": "546201",
      "postDate": "06/06/2019 10:59:04",
      "content": "<p>Sorry for being a jerk. But why does it matter? Will you approach the solution differently depending upon how was it generated?</p>",
      "rawMarkdown": "Sorry for being a jerk. But why does it matter? Will you approach the solution differently depending upon how was it generated?",
      "votes": null
    },
    {
      "id": "546209",
      "postDate": "06/06/2019 11:05:24",
      "content": "<p>if its from simulation we can assume zero error theoretically if somebody re-do the simu exactly.</p>",
      "rawMarkdown": "if its from simulation we can assume zero error theoretically if somebody re-do the simu exactly.",
      "votes": null
    },
    {
      "id": "546210",
      "postDate": "06/06/2019 11:07:17",
      "content": "<p>the error seems sooo low. If you look at target stats: mean~16 std~35. And the model has MAE=0.5 (~ -1.0 lb score), which means almost perfect model.</p>",
      "rawMarkdown": "the error seems sooo low. If you look at target stats: mean~16 std~35. And the model has MAE=0.5 (~ -1.0 lb score), which means almost perfect model.",
      "votes": null
    },
    {
      "id": "546288",
      "postDate": "06/06/2019 12:39:02",
      "content": "<p>Thanks a million. Learnt something new today. 👍 </p>",
      "rawMarkdown": "Thanks a million. Learnt something new today. 👍",
      "votes": null
    },
    {
      "id": "546302",
      "postDate": "06/06/2019 13:00:04",
      "content": "<p>The dataset is from theoretical computation, not experiments. In fact, the values don't agree with experiment.</p>",
      "rawMarkdown": "The dataset is from theoretical computation, not experiments. In fact, the values don't agree with experiment.",
      "votes": null
    },
    {
      "id": "548568",
      "postDate": "06/09/2019 15:28:02",
      "content": "<p>you mean all data include dev and test is from theoretical computation? It is interesting.</p>",
      "rawMarkdown": "you mean all data include dev and test is from theoretical computation? It is interesting.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 546201,
      "author_name": "cyberia",
      "author_url": "",
      "post_date": "06/06/2019 10:59:04",
      "content": "<p>Sorry for being a jerk. But why does it matter? Will you approach the solution differently depending upon how was it generated?</p>",
      "votes": null,
      "replies": [
        {
          "id": 546209,
          "author_name": "mjahrer",
          "author_url": "",
          "post_date": "06/06/2019 11:05:24",
          "content": "<p>if its from simulation we can assume zero error theoretically if somebody re-do the simu exactly.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 546210,
          "author_name": "mjahrer",
          "author_url": "",
          "post_date": "06/06/2019 11:07:17",
          "content": "<p>the error seems sooo low. If you look at target stats: mean~16 std~35. And the model has MAE=0.5 (~ -1.0 lb score), which means almost perfect model.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 546288,
          "author_name": "cyberia",
          "author_url": "",
          "post_date": "06/06/2019 12:39:02",
          "content": "<p>Thanks a million. Learnt something new today. 👍 </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 546302,
      "author_name": "phucco",
      "author_url": "",
      "post_date": "06/06/2019 13:00:04",
      "content": "<p>The dataset is from theoretical computation, not experiments. In fact, the values don't agree with experiment.</p>",
      "votes": null,
      "replies": [
        {
          "id": 548568,
          "author_name": "al3chen",
          "author_url": "",
          "post_date": "06/09/2019 15:28:02",
          "content": "<p>you mean all data include dev and test is from theoretical computation? It is interesting.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "546177": "I just read the description and checked discussion threads. But cannot find the answer. So it this dataset from a chemistry simulation/theory/formulas or are they from experimental results?",
    "546201": "Sorry for being a jerk. But why does it matter? Will you approach the solution differently depending upon how was it generated?",
    "546209": "if its from simulation we can assume zero error theoretically if somebody re-do the simu exactly.",
    "546210": "the error seems sooo low. If you look at target stats: mean~16 std~35. And the model has MAE=0.5 (~ -1.0 lb score), which means almost perfect model.",
    "546288": "Thanks a million. Learnt something new today. 👍",
    "546302": "The dataset is from theoretical computation, not experiments. In fact, the values don't agree with experiment.",
    "548568": "you mean all data include dev and test is from theoretical computation? It is interesting."
  },
  "source": "meta"
}