I have 2 doubts related to the dataset
\n\n\ndifficult_negative_case - True if the case was unusually difficult. Only provided for train
\n
\n How should we consider this column in terms of probability of cancer?
Should we take this column as \"there is no cancer but the case was complicated\" or should we take it as \"there is no cancer, but our conclusion might be wrong. Hence, this patient might have cancer?\".
\nSo data with \"True\" complicated values should be treated as \"Cancer Patients\"?
\n\n\n\nBIRADS - 0 if the breast required follow-up, 1 if the breast was rated as negative for cancer, and 2 if the breast was rated as normal. Only provided for train.
\n
. 2nd doubt answered here by David, thanks.
", "messages": [ { "id": 2051383, "postDate": "2022-12-01T11:29:46.490Z", "content": "From my understanding, you should interpret it as \"a case that is actually negative but it is hard to categorise it as negative\"
", "rawMarkdown": "From my understanding, you should interpret it as \"a case that is actually negative but it is hard to categorise it as negative\"", "votes": 4, "replies": [ { "id": 2051941, "postDate": "2022-12-01T18:38:13.847Z", "content": "This is correct.
", "rawMarkdown": "This is correct.", "votes": 2 }, { "id": 2052220, "postDate": "2022-12-02T02:45:06.057Z", "content": "Okay, thank you.
", "rawMarkdown": "Okay, thank you. " }, { "id": 2052586, "postDate": "2022-12-02T10:13:57.597Z", "content": "I still have a confusion. When the cancer is 1, does it mean that it actually gets a cancer or there is a probability of false positive? I mean whether the lable \"cancer\" represents the actual situation?
", "rawMarkdown": "I still have a confusion. When the cancer is 1, does it mean that it actually gets a cancer or there is a probability of false positive? I mean whether the lable \"cancer\" represents the actual situation?", "votes": 1 }, { "id": 2052832, "postDate": "2022-12-02T14:42:24.070Z", "content": "That's what I was confused about. Same goes for \"difficult_negative_case\", like is there a probability that it was mis-diagnosed. But, it seems like that data is taken on \"As-Is\" basis.
", "rawMarkdown": "That's what I was confused about. Same goes for \"difficult_negative_case\", like is there a probability that it was mis-diagnosed. But, it seems like that data is taken on \"As-Is\" basis." }, { "id": 2053060, "postDate": "2022-12-02T19:06:18.087Z", "content": "@hrlmmm I believe the positive cancer label was verified with biopsies.
", "rawMarkdown": "@hrlmmm I believe the positive cancer label was verified with biopsies." }, { "id": 2053287, "postDate": "2022-12-03T04:42:41.810Z", "content": "Do you means that all the \"TRUE\" cases have no cancer in fact, whether they had a biopsy or not…When the data with \"True\" complicated values, it just means that it's difficult for the doctor to judge? and those negative cancer labels sitll represent the actual situation, without false negative case.
", "rawMarkdown": "Do you means that all the \"TRUE\" cases have no cancer in fact, whether they had a biopsy or not...When the data with \"True\" complicated values, it just means that it's difficult for the doctor to judge? and those negative cancer labels sitll represent the actual situation, without false negative case.", "votes": 1 }, { "id": 2056075, "postDate": "2022-12-05T17:50:48.797Z", "content": "If a difficult case did not get a biopsy, then the cancer would have been ruled out by a follow up diagnostic mammogram. I'm very much not an expert in this domain, but I believe those use higher doses of radiation to generate higher resolution and/or 3D images.
\nFalse negatives are unlikely but possible; that's just the nature of diagnosing cancer.
", "rawMarkdown": "If a difficult case did not get a biopsy, then the cancer would have been ruled out by a follow up diagnostic mammogram. I'm very much not an expert in this domain, but I believe those use higher doses of radiation to generate higher resolution and/or 3D images.\n\nFalse negatives are unlikely but possible; that's just the nature of diagnosing cancer." }, { "id": 2056797, "postDate": "2022-12-06T12:47:55.850Z", "content": "Thanks for answer !!
", "rawMarkdown": "Thanks for answer !! " } ] }, { "id": 2048526, "postDate": "2022-11-29T14:12:15.037Z", "content": "I have 2 doubts related to the dataset
\n\n\ndifficult_negative_case - True if the case was unusually difficult. Only provided for train
\n
\n How should we consider this column in terms of probability of cancer?
Should we take this column as \"there is no cancer but the case was complicated\" or should we take it as \"there is no cancer, but our conclusion might be wrong. Hence, this patient might have cancer?\".
\nSo data with \"True\" complicated values should be treated as \"Cancer Patients\"?
\n\n\n\nBIRADS - 0 if the breast required follow-up, 1 if the breast was rated as negative for cancer, and 2 if the breast was rated as normal. Only provided for train.
\n
. 2nd doubt answered here by David, thanks.
", "rawMarkdown": "I have 2 doubts related to the dataset\n\n- 1st doubt\n> difficult_negative_case - True if the case was unusually difficult. Only provided for train\nHow should we consider this column in terms of probability of cancer? \n\nShould we take this column as \"there is no cancer but the case was complicated\" or should we take it as \"there is no cancer, but our conclusion might be wrong. Hence, this patient might have cancer?\". \n\nSo data with \"True\" complicated values should be treated as \"Cancer Patients\"?\n\n~~- 2nd doubt~~\n> BIRADS - 0 if the breast required follow-up, 1 if the breast was rated as negative for cancer, and 2 if the breast was rated as normal. Only provided for train.\n\n~~\nHow's negative different than normal? Isn't \"1\" negative is equal to \"2\" (normal)? If a person is normal doesn't it mean that the they are negative as well?~~. 2nd doubt answered [here](https://www.kaggle.com/competitions/rsna-breast-cancer-detection/discussion/369135#2048552) by David, thanks.\n", "votes": 4 } ], "comments": [ { "id": 2051383, "author_name": "AssskyRui", "author_url": "", "post_date": "2022-12-01T11:29:46.490000", "content": "From my understanding, you should interpret it as \"a case that is actually negative but it is hard to categorise it as negative\"
", "votes": 4, "replies": [ { "id": 2051941, "author_name": "Sohier Dane", "author_url": "", "post_date": "2022-12-01T18:38:13.847000", "content": "This is correct.
", "votes": 2, "replies": [] }, { "id": 2052220, "author_name": "Hey24sheep", "author_url": "", "post_date": "2022-12-02T02:45:06.057000", "content": "Okay, thank you.
", "votes": 0, "replies": [] }, { "id": 2052586, "author_name": "hrlmmm", "author_url": "", "post_date": "2022-12-02T10:13:57.597000", "content": "I still have a confusion. When the cancer is 1, does it mean that it actually gets a cancer or there is a probability of false positive? I mean whether the lable \"cancer\" represents the actual situation?
", "votes": 1, "replies": [] }, { "id": 2052832, "author_name": "Hey24sheep", "author_url": "", "post_date": "2022-12-02T14:42:24.070000", "content": "That's what I was confused about. Same goes for \"difficult_negative_case\", like is there a probability that it was mis-diagnosed. But, it seems like that data is taken on \"As-Is\" basis.
", "votes": 0, "replies": [] }, { "id": 2053060, "author_name": "Sohier Dane", "author_url": "", "post_date": "2022-12-02T19:06:18.087000", "content": "@hrlmmm I believe the positive cancer label was verified with biopsies.
", "votes": 0, "replies": [] }, { "id": 2053287, "author_name": "hrlmmm", "author_url": "", "post_date": "2022-12-03T04:42:41.810000", "content": "Do you means that all the \"TRUE\" cases have no cancer in fact, whether they had a biopsy or not…When the data with \"True\" complicated values, it just means that it's difficult for the doctor to judge? and those negative cancer labels sitll represent the actual situation, without false negative case.
", "votes": 1, "replies": [] }, { "id": 2056075, "author_name": "Sohier Dane", "author_url": "", "post_date": "2022-12-05T17:50:48.797000", "content": "If a difficult case did not get a biopsy, then the cancer would have been ruled out by a follow up diagnostic mammogram. I'm very much not an expert in this domain, but I believe those use higher doses of radiation to generate higher resolution and/or 3D images.
\nFalse negatives are unlikely but possible; that's just the nature of diagnosing cancer.
", "votes": 0, "replies": [] }, { "id": 2056797, "author_name": "hrlmmm", "author_url": "", "post_date": "2022-12-06T12:47:55.850000", "content": "Thanks for answer !!
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "2051383": "From my understanding, you should interpret it as \"a case that is actually negative but it is hard to categorise it as negative\"", "2048526": "I have 2 doubts related to the dataset\n\n- 1st doubt\n> difficult_negative_case - True if the case was unusually difficult. Only provided for train\nHow should we consider this column in terms of probability of cancer? \n\nShould we take this column as \"there is no cancer but the case was complicated\" or should we take it as \"there is no cancer, but our conclusion might be wrong. Hence, this patient might have cancer?\". \n\nSo data with \"True\" complicated values should be treated as \"Cancer Patients\"?\n\n~~- 2nd doubt~~\n> BIRADS - 0 if the breast required follow-up, 1 if the breast was rated as negative for cancer, and 2 if the breast was rated as normal. Only provided for train.\n\n~~\nHow's negative different than normal? Isn't \"1\" negative is equal to \"2\" (normal)? If a person is normal doesn't it mean that the they are negative as well?~~. 2nd doubt answered [here](https://www.kaggle.com/competitions/rsna-breast-cancer-detection/discussion/369135#2048552) by David, thanks.\n" } }