{
  "id": 400048,
  "title": "What are we supposed to submit? ",
  "url": "/competitions/tlvmc-parkinsons-freezing-gait-prediction/discussion/400048",
  "author_name": "",
  "post_date": "2023-04-06T16:07:16.143106400Z",
  "votes": 1,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Hi,</p>\n<p>Are we supposed to submit a .csv file for the</p>\n<ul>\n<li>tdcsfog <br>\nor</li>\n<li>defog<br>\ntest data set? Or both? </li>\n</ul>\n<p>How is it possible for hidden data to be used for scoring, if we only submit a .csv file for a test data subset? </p>\n<p>Best,<br>\nFelix</p>",
  "messages": [
    {
      "id": "2212227",
      "postDate": "04/06/2023 16:07:16",
      "content": "<p>Hi,</p>\n<p>Are we supposed to submit a .csv file for the</p>\n<ul>\n<li>tdcsfog <br>\nor</li>\n<li>defog<br>\ntest data set? Or both? </li>\n</ul>\n<p>How is it possible for hidden data to be used for scoring, if we only submit a .csv file for a test data subset? </p>\n<p>Best,<br>\nFelix</p>",
      "rawMarkdown": "Hi,\n\nAre we supposed to submit a .csv file for the\n- tdcsfog \nor\n- defog\ntest data set? Or both? \n\nHow is it possible for hidden data to be used for scoring, if we only submit a .csv file for a test data subset? \n\nBest,\nFelix",
      "votes": null
    },
    {
      "id": "2212293",
      "postDate": "04/06/2023 17:10:18",
      "content": "<p>Hi Felix, the sample submission file shows results for both a TDCS test set and a Defog test set. Your notebook should be general enough that upon submission, a hidden test data set can be read in, predicted, and scored. In other words, we should avoid reading in the visible test data by name and instead read all test files in the TDCS and Defog folders. The folders will contain many more files upon submission.</p>\n<p>Edit: this is all submitted in one file.</p>",
      "rawMarkdown": "Hi Felix, the sample submission file shows results for both a TDCS test set and a Defog test set. Your notebook should be general enough that upon submission, a hidden test data set can be read in, predicted, and scored. In other words, we should avoid reading in the visible test data by name and instead read all test files in the TDCS and Defog folders. The folders will contain many more files upon submission.\n\nEdit: this is all submitted in one file.",
      "votes": null
    },
    {
      "id": "2212365",
      "postDate": "04/06/2023 18:08:13",
      "content": "<p>Hi Austin, </p>\n<p>Thank you very much for the explanation!<br>\nThat means that</p>\n<ol>\n<li>I need to assume that the \\test directory contains more test data sets for tdcsfog and defog for the final scoring? Thus, defining the code so that it is able to read more than one data set per directory</li>\n<li>I will apply my model both to the tdcsfog and defog test set, concatenate the predictions and then save it as the submission.csv.</li>\n</ol>\n<p>I that correct?</p>\n<p>Best,<br>\nFelix  </p>",
      "rawMarkdown": "Hi Austin, \n\nThank you very much for the explanation!\nThat means that\n1. I need to assume that the \\test directory contains more test data sets for tdcsfog and defog for the final scoring? Thus, defining the code so that it is able to read more than one data set per directory\n2. I will apply my model both to the tdcsfog and defog test set, concatenate the predictions and then save it as the submission.csv.\n\nI that correct?\n\nBest,\nFelix",
      "votes": null
    },
    {
      "id": "2212401",
      "postDate": "04/06/2023 18:35:54",
      "content": "<p>I think so — with the caveat that the sampling rates are different for Defog data and TDCS data (100 and 128 Hz respectively) as are the conditions under which the data was collected, so I think some folks are using two separate models, while others preprocess the data to account for these differences.  In any event, there is only one submission file so you’ll have to concatenate the two types eventually.</p>",
      "rawMarkdown": "I think so — with the caveat that the sampling rates are different for Defog data and TDCS data (100 and 128 Hz respectively) as are the conditions under which the data was collected, so I think some folks are using two separate models, while others preprocess the data to account for these differences.  In any event, there is only one submission file so you’ll have to concatenate the two types eventually.",
      "votes": null
    },
    {
      "id": "2212455",
      "postDate": "04/06/2023 19:16:27",
      "content": "<p>Thank you, that helped a lot!</p>",
      "rawMarkdown": "Thank you, that helped a lot!",
      "votes": null
    },
    {
      "id": "2214708",
      "postDate": "04/08/2023 17:23:57",
      "content": "<p>Not only the sampling rate but also the units (g vs m/s2)</p>",
      "rawMarkdown": "Not only the sampling rate but also the units (g vs m/s2)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2212293,
      "author_name": "austinhinkel",
      "author_url": "",
      "post_date": "04/06/2023 17:10:18",
      "content": "<p>Hi Felix, the sample submission file shows results for both a TDCS test set and a Defog test set. Your notebook should be general enough that upon submission, a hidden test data set can be read in, predicted, and scored. In other words, we should avoid reading in the visible test data by name and instead read all test files in the TDCS and Defog folders. The folders will contain many more files upon submission.</p>\n<p>Edit: this is all submitted in one file.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2212365,
      "author_name": "frichter",
      "author_url": "",
      "post_date": "04/06/2023 18:08:13",
      "content": "<p>Hi Austin, </p>\n<p>Thank you very much for the explanation!<br>\nThat means that</p>\n<ol>\n<li>I need to assume that the \\test directory contains more test data sets for tdcsfog and defog for the final scoring? Thus, defining the code so that it is able to read more than one data set per directory</li>\n<li>I will apply my model both to the tdcsfog and defog test set, concatenate the predictions and then save it as the submission.csv.</li>\n</ol>\n<p>I that correct?</p>\n<p>Best,<br>\nFelix  </p>",
      "votes": null,
      "replies": [
        {
          "id": 2212401,
          "author_name": "austinhinkel",
          "author_url": "",
          "post_date": "04/06/2023 18:35:54",
          "content": "<p>I think so — with the caveat that the sampling rates are different for Defog data and TDCS data (100 and 128 Hz respectively) as are the conditions under which the data was collected, so I think some folks are using two separate models, while others preprocess the data to account for these differences.  In any event, there is only one submission file so you’ll have to concatenate the two types eventually.</p>",
          "votes": null,
          "replies": [
            {
              "id": 2212455,
              "author_name": "frichter",
              "author_url": "",
              "post_date": "04/06/2023 19:16:27",
              "content": "<p>Thank you, that helped a lot!</p>",
              "votes": null,
              "replies": []
            },
            {
              "id": 2214708,
              "author_name": "albertoannoni",
              "author_url": "",
              "post_date": "04/08/2023 17:23:57",
              "content": "<p>Not only the sampling rate but also the units (g vs m/s2)</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2212227": "Hi,\n\nAre we supposed to submit a .csv file for the\n- tdcsfog \nor\n- defog\ntest data set? Or both? \n\nHow is it possible for hidden data to be used for scoring, if we only submit a .csv file for a test data subset? \n\nBest,\nFelix",
    "2212293": "Hi Felix, the sample submission file shows results for both a TDCS test set and a Defog test set. Your notebook should be general enough that upon submission, a hidden test data set can be read in, predicted, and scored. In other words, we should avoid reading in the visible test data by name and instead read all test files in the TDCS and Defog folders. The folders will contain many more files upon submission.\n\nEdit: this is all submitted in one file.",
    "2212365": "Hi Austin, \n\nThank you very much for the explanation!\nThat means that\n1. I need to assume that the \\test directory contains more test data sets for tdcsfog and defog for the final scoring? Thus, defining the code so that it is able to read more than one data set per directory\n2. I will apply my model both to the tdcsfog and defog test set, concatenate the predictions and then save it as the submission.csv.\n\nI that correct?\n\nBest,\nFelix",
    "2212401": "I think so — with the caveat that the sampling rates are different for Defog data and TDCS data (100 and 128 Hz respectively) as are the conditions under which the data was collected, so I think some folks are using two separate models, while others preprocess the data to account for these differences.  In any event, there is only one submission file so you’ll have to concatenate the two types eventually.",
    "2212455": "Thank you, that helped a lot!",
    "2214708": "Not only the sampling rate but also the units (g vs m/s2)"
  },
  "source": "meta"
}