{
  "id": 471521,
  "title": "EEG_SUB_ID Mapping",
  "url": "/competitions/hms-harmful-brain-activity-classification/discussion/471521",
  "author_name": "",
  "post_date": "2024-01-28T16:27:52.124195800Z",
  "votes": 7,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi guys, <br>\nfor all of you who want to map the eeg_sub_id to the specific part of the eeg date, here is the solution :) </p>\n<p>Usually the eeg data for an eeg_id contains 10k rows because we have 50 seconds x 200 records per second. If we have multiple entries for our eeg_id in the metadate, it will have  eeg_sub_ids. Next to the sub_id you can find the time (in seconds) when this subrecord starts. If we want to find the starting row, we have to use the mentioned formula minus one because the index starts with 0.</p>\n<p>So let's imagine our last sub record starts after 20 seconds. To find the starting row, we multiply 20 by 200 and subtract one. So we know our first line for this sub id is 3999. Our last row for this sub_id is always 10k lines later which means row 13999.</p>\n<p>I hope it will help you for EDAs and maybe for more training data if needed</p>",
  "messages": [
    {
      "id": "2624180",
      "postDate": "01/28/2024 16:27:52",
      "content": "<p>Hi guys, <br>\nfor all of you who want to map the eeg_sub_id to the specific part of the eeg date, here is the solution :) </p>\n<p>Usually the eeg data for an eeg_id contains 10k rows because we have 50 seconds x 200 records per second. If we have multiple entries for our eeg_id in the metadate, it will have  eeg_sub_ids. Next to the sub_id you can find the time (in seconds) when this subrecord starts. If we want to find the starting row, we have to use the mentioned formula minus one because the index starts with 0.</p>\n<p>So let's imagine our last sub record starts after 20 seconds. To find the starting row, we multiply 20 by 200 and subtract one. So we know our first line for this sub id is 3999. Our last row for this sub_id is always 10k lines later which means row 13999.</p>\n<p>I hope it will help you for EDAs and maybe for more training data if needed</p>",
      "rawMarkdown": "Hi guys, \nfor all of you who want to map the eeg_sub_id to the specific part of the eeg date, here is the solution :) \n\nUsually the eeg data for an eeg_id contains 10k rows because we have 50 seconds x 200 records per second. If we have multiple entries for our eeg_id in the metadate, it will have  eeg_sub_ids. Next to the sub_id you can find the time (in seconds) when this subrecord starts. If we want to find the starting row, we have to use the mentioned formula minus one because the index starts with 0.\n\nSo let's imagine our last sub record starts after 20 seconds. To find the starting row, we multiply 20 by 200 and subtract one. So we know our first line for this sub id is 3999. Our last row for this sub_id is always 10k lines later which means row 13999.\n\nI hope it will help you for EDAs and maybe for more training data if needed",
      "votes": null
    },
    {
      "id": "2626022",
      "postDate": "01/29/2024 18:33:21",
      "content": "<p>Thanks, This was really helpfull!!</p>\n<p>Can you please confirm for spectogram also, this is same? </p>\n<p>Without Offset Information shape:</p>\n<p>EEG: 10000 -- 50 seconds * 200<br>\nSpectogram: 300 -- 30 * 10 Minutes</p>\n<p>With 40 Second offsets EEG Shape<br>\nEEG: 18000 -- 90(50 + 40) seconds * 200<br>\nSpectogram: 320 -- 10.67 * 30</p>",
      "rawMarkdown": "Thanks, This was really helpfull!!\n\nCan you please confirm for spectogram also, this is same? \n\nWithout Offset Information shape:\n\nEEG: 10000 -- 50 seconds * 200\nSpectogram: 300 -- 30 * 10 Minutes\n\nWith 40 Second offsets EEG Shape\nEEG: 18000 -- 90(50 + 40) seconds * 200\nSpectogram: 320 -- 10.67 * 30",
      "votes": null
    },
    {
      "id": "2642582",
      "postDate": "02/08/2024 09:08:31",
      "content": "<p>should be simmilar </p>",
      "rawMarkdown": "should be simmilar",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2626022,
      "author_name": "wasimmadha",
      "author_url": "",
      "post_date": "01/29/2024 18:33:21",
      "content": "<p>Thanks, This was really helpfull!!</p>\n<p>Can you please confirm for spectogram also, this is same? </p>\n<p>Without Offset Information shape:</p>\n<p>EEG: 10000 -- 50 seconds * 200<br>\nSpectogram: 300 -- 30 * 10 Minutes</p>\n<p>With 40 Second offsets EEG Shape<br>\nEEG: 18000 -- 90(50 + 40) seconds * 200<br>\nSpectogram: 320 -- 10.67 * 30</p>",
      "votes": null,
      "replies": [
        {
          "id": 2642582,
          "author_name": "oles04",
          "author_url": "",
          "post_date": "02/08/2024 09:08:31",
          "content": "<p>should be simmilar </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2624180": "Hi guys, \nfor all of you who want to map the eeg_sub_id to the specific part of the eeg date, here is the solution :) \n\nUsually the eeg data for an eeg_id contains 10k rows because we have 50 seconds x 200 records per second. If we have multiple entries for our eeg_id in the metadate, it will have  eeg_sub_ids. Next to the sub_id you can find the time (in seconds) when this subrecord starts. If we want to find the starting row, we have to use the mentioned formula minus one because the index starts with 0.\n\nSo let's imagine our last sub record starts after 20 seconds. To find the starting row, we multiply 20 by 200 and subtract one. So we know our first line for this sub id is 3999. Our last row for this sub_id is always 10k lines later which means row 13999.\n\nI hope it will help you for EDAs and maybe for more training data if needed",
    "2626022": "Thanks, This was really helpfull!!\n\nCan you please confirm for spectogram also, this is same? \n\nWithout Offset Information shape:\n\nEEG: 10000 -- 50 seconds * 200\nSpectogram: 300 -- 30 * 10 Minutes\n\nWith 40 Second offsets EEG Shape\nEEG: 18000 -- 90(50 + 40) seconds * 200\nSpectogram: 320 -- 10.67 * 30",
    "2642582": "should be simmilar"
  },
  "source": "meta"
}