{ "id": 177061, "title": "What does this mean by \"final three FVC measurements\" ? ", "url": "/competitions/osic-pulmonary-fibrosis-progression/discussion/177061", "author_name": "", "post_date": "2020-08-24T16:19:11.554800Z", "votes": 9, "comment_count": 7, "views": 0, "content": "

Hi All,

\n

Could someone please help me understand, what is \"final three FVC measurements\" ?

\n

In the Data page of the competition, it is mentioned that You are asked to predict the final three FVC measurements for each patient, as well as a confidence value in your prediction.

\n

Thank you.

", "messages": [ { "id": "983847", "postDate": "08/24/2020 16:19:11", "content": "

Hi All,

\n

Could someone please help me understand, what is \"final three FVC measurements\" ?

\n

In the Data page of the competition, it is mentioned that You are asked to predict the final three FVC measurements for each patient, as well as a confidence value in your prediction.

\n

Thank you.

", "rawMarkdown": "Hi All,\n\nCould someone please help me understand, what is **\"final three FVC measurements\"** ?\n\nIn the Data page of the competition, it is mentioned that **You are asked to predict the final three FVC measurements for each patient, as well as a confidence value in your prediction.**\n\n\nThank you.", "votes": null }, { "id": "983853", "postDate": "08/24/2020 16:23:52", "content": "

For FVC Check this

\n

and Confidence is Standard deviation

", "rawMarkdown": "For FVC [Check this](https://lunginstitute.com/blog/fev1-and-fvc/#:~:text=The%20forced%20vital%20capacity%20(FVC,the%20severity%20of%20the%20condition)\n\nand Confidence is Standard deviation", "votes": null }, { "id": "983871", "postDate": "08/24/2020 16:43:31", "content": "

Hi,
\nIn your submission, you provide predictions for every week within the range (-12 to 133). However, you will only be evaluated on the final three FVC measurements for this patient.
\nFor example: if this patient had FVC measurements in weeks: 0,5,10,15,20,25,35: your predictions will be evaluated in the weeks 20, 25, and 35 only. You don't know these weeks for each patient, thus you provide predictions for all weeks and at evaluation time only the last three visits are selected.
\nHope this helps.

", "rawMarkdown": "Hi,\nIn your submission, you provide predictions for every week within the range (-12 to 133). However, you will only be evaluated on the final three FVC measurements for this patient.\nFor example: if this patient had FVC measurements in weeks: 0,5,10,15,20,25,35: your predictions will be evaluated in the weeks 20, 25, and 35 only. You don't know these weeks for each patient, thus you provide predictions for all weeks and at evaluation time only the last three visits are selected. \nHope this helps.", "votes": null }, { "id": "983888", "postDate": "08/24/2020 17:00:49", "content": "

Hey @ahmedhshahin I have two questions for you:

\n
    \n

  1. The weeks range should be -12 to 133 right? Is any week not from this range possible? Should I write any code to filter only these weeks data. Because in the evaluation section of competition overview, this is written there : you are asked to predict every patient's FVC measurement for every possible week.

    2. \n

  2. Can it be possible that features of meta data of DICOM images may not be same in public and private data? Because I am getting the well known Submission Scoring Error when I use meta for modelling, while not in the case of using only original features.

    3. \n
\n

I suspect that the error can be possible from the above two problems.

\n

Please help in clearing my query!

", "rawMarkdown": "Hey @ahmedhshahin I have two questions for you:\n\n1. The weeks range should be -12 to 133 right? Is any week not from this range possible? Should I write any code to filter only these weeks data. Because in the evaluation section of competition overview, this is written there : `you are asked to predict every patient's FVC measurement for every possible week.`\n\n2. Can it be possible that features of meta data of DICOM images may not be same in public and private data? Because I am getting the well known `Submission Scoring Error` when I use meta for modelling, while not in the case of using only original features.\n\nI suspect that the error can be possible from the above two problems.\n\nPlease help in clearing my query!", "votes": null }, { "id": "983923", "postDate": "08/24/2020 17:28:48", "content": "
    \n
  1. No, actually this range is selected to cover all possible weeks in the test set.
    2. \n
  2. We have discussed this in other topics and concluded that all needed tags exist in the test set, also all test set CT volumes are readable without problems.
    3. \n
\n

This could be something with your submission code (the part that generates the CSV file), I would suggest going through one of the popular notebooks and checking their submission part to see if you are doing something wrong.

", "rawMarkdown": "1. No, actually this range is selected to cover all possible weeks in the test set.\n2. We have discussed this in other topics and concluded that all needed tags exist in the test set, also all test set CT volumes are readable without problems.\n\nThis could be something with your submission code (the part that generates the CSV file), I would suggest going through one of the popular notebooks and checking their submission part to see if you are doing something wrong.", "votes": null }, { "id": "983929", "postDate": "08/24/2020 17:34:07", "content": "

Sure thanks, @ahmedhshahin !

", "rawMarkdown": "Sure thanks, @ahmedhshahin !", "votes": null }, { "id": "984331", "postDate": "08/25/2020 03:36:54", "content": "

Thank you @ahmedhshahin
\nThis answers my question.

", "rawMarkdown": "Thank you @ahmedhshahin\nThis answers my question.", "votes": null }, { "id": "986192", "postDate": "08/26/2020 09:22:07", "content": "

Thank you @aman2000jaiswal

", "rawMarkdown": "Thank you @aman2000jaiswal", "votes": null } ], "comments": [ { "id": 983853, "author_name": "aman2000jaiswal", "author_url": "", "post_date": "08/24/2020 16:23:52", "content": "

For FVC Check this

\n

and Confidence is Standard deviation

", "votes": null, "replies": [ { "id": 986192, "author_name": "ashutosh3060", "author_url": "", "post_date": "08/26/2020 09:22:07", "content": "

Thank you @aman2000jaiswal

", "votes": null, "replies": [] } ] }, { "id": 983871, "author_name": "ahmedhshahin", "author_url": "", "post_date": "08/24/2020 16:43:31", "content": "

Hi,
\nIn your submission, you provide predictions for every week within the range (-12 to 133). However, you will only be evaluated on the final three FVC measurements for this patient.
\nFor example: if this patient had FVC measurements in weeks: 0,5,10,15,20,25,35: your predictions will be evaluated in the weeks 20, 25, and 35 only. You don't know these weeks for each patient, thus you provide predictions for all weeks and at evaluation time only the last three visits are selected.
\nHope this helps.

", "votes": null, "replies": [ { "id": 983888, "author_name": "sarques", "author_url": "", "post_date": "08/24/2020 17:00:49", "content": "

Hey @ahmedhshahin I have two questions for you:

\n
    \n

  1. The weeks range should be -12 to 133 right? Is any week not from this range possible? Should I write any code to filter only these weeks data. Because in the evaluation section of competition overview, this is written there : you are asked to predict every patient's FVC measurement for every possible week.

    2. \n

  2. Can it be possible that features of meta data of DICOM images may not be same in public and private data? Because I am getting the well known Submission Scoring Error when I use meta for modelling, while not in the case of using only original features.

    3. \n
\n

I suspect that the error can be possible from the above two problems.

\n

Please help in clearing my query!

", "votes": null, "replies": [] }, { "id": 983923, "author_name": "ahmedhshahin", "author_url": "", "post_date": "08/24/2020 17:28:48", "content": "
    \n
  1. No, actually this range is selected to cover all possible weeks in the test set.
    2. \n
  2. We have discussed this in other topics and concluded that all needed tags exist in the test set, also all test set CT volumes are readable without problems.
    3. \n
\n

This could be something with your submission code (the part that generates the CSV file), I would suggest going through one of the popular notebooks and checking their submission part to see if you are doing something wrong.

", "votes": null, "replies": [] }, { "id": 983929, "author_name": "sarques", "author_url": "", "post_date": "08/24/2020 17:34:07", "content": "

Sure thanks, @ahmedhshahin !

", "votes": null, "replies": [] }, { "id": 984331, "author_name": "ashutosh3060", "author_url": "", "post_date": "08/25/2020 03:36:54", "content": "

Thank you @ahmedhshahin
\nThis answers my question.

", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "983847": "Hi All,\n\nCould someone please help me understand, what is **\"final three FVC measurements\"** ?\n\nIn the Data page of the competition, it is mentioned that **You are asked to predict the final three FVC measurements for each patient, as well as a confidence value in your prediction.**\n\n\nThank you.", "983853": "For FVC [Check this](https://lunginstitute.com/blog/fev1-and-fvc/#:~:text=The%20forced%20vital%20capacity%20(FVC,the%20severity%20of%20the%20condition)\n\nand Confidence is Standard deviation", "983871": "Hi,\nIn your submission, you provide predictions for every week within the range (-12 to 133). However, you will only be evaluated on the final three FVC measurements for this patient.\nFor example: if this patient had FVC measurements in weeks: 0,5,10,15,20,25,35: your predictions will be evaluated in the weeks 20, 25, and 35 only. You don't know these weeks for each patient, thus you provide predictions for all weeks and at evaluation time only the last three visits are selected. \nHope this helps.", "983888": "Hey @ahmedhshahin I have two questions for you:\n\n1. The weeks range should be -12 to 133 right? Is any week not from this range possible? Should I write any code to filter only these weeks data. Because in the evaluation section of competition overview, this is written there : `you are asked to predict every patient's FVC measurement for every possible week.`\n\n2. Can it be possible that features of meta data of DICOM images may not be same in public and private data? Because I am getting the well known `Submission Scoring Error` when I use meta for modelling, while not in the case of using only original features.\n\nI suspect that the error can be possible from the above two problems.\n\nPlease help in clearing my query!", "983923": "1. No, actually this range is selected to cover all possible weeks in the test set.\n2. We have discussed this in other topics and concluded that all needed tags exist in the test set, also all test set CT volumes are readable without problems.\n\nThis could be something with your submission code (the part that generates the CSV file), I would suggest going through one of the popular notebooks and checking their submission part to see if you are doing something wrong.", "983929": "Sure thanks, @ahmedhshahin !", "984331": "Thank you @ahmedhshahin\nThis answers my question.", "986192": "Thank you @aman2000jaiswal" }, "source": "meta" }