What and how many are included in the label considered The Other?
\nIn my opinion, the dataset seems to cover the four phenotypes commonly seen in ovarian cancer (although there are two for serous carcinoma) as intended.
\nIf the Test data is sampled normally, ignoring The Other label would produce a high enough score, but I don't think that would be consistent with the questioner's intention.
\nI would like to hear the opinions of all participants.
", "messages": [ { "id": 2518041, "postDate": "2023-11-09T01:55:43.663Z", "content": "What and how many are included in the label considered The Other?
\nIn my opinion, the dataset seems to cover the four phenotypes commonly seen in ovarian cancer (although there are two for serous carcinoma) as intended.
\nIf the Test data is sampled normally, ignoring The Other label would produce a high enough score, but I don't think that would be consistent with the questioner's intention.
\nI would like to hear the opinions of all participants.
", "rawMarkdown": "What and how many are included in the label considered The Other?\n\nIn my opinion, the dataset seems to cover the four phenotypes commonly seen in ovarian cancer (although there are two for serous carcinoma) as intended.\n\nIf the Test data is sampled normally, ignoring The Other label would produce a high enough score, but I don't think that would be consistent with the questioner's intention.\n\nI would like to hear the opinions of all participants.", "votes": 13 }, { "id": 2519566, "postDate": "2023-11-10T06:58:34.097Z", "content": "Based on the metric selection, Other class seems to be equally important from the clinical perspective. Even if you predict all of the samples correctly and ignore Other, you can reach 83.33% balanced accuracy. However someone with a slightly less accurate model that can capture couple Other cases might score better than you.
", "rawMarkdown": "Based on the metric selection, Other class seems to be equally important from the clinical perspective. Even if you predict all of the samples correctly and ignore Other, you can reach 83.33% balanced accuracy. However someone with a slightly less accurate model that can capture couple Other cases might score better than you. ", "votes": 3, "replies": [ { "id": 2519597, "postDate": "2023-11-10T07:23:07.600Z", "content": "Thanks for your valuable comments!
\nAs a biochemistry major myself, the choice of dataset is a bit strange in terms of practicality.
I wonder if the second half of this competition will be a battle to see how well we can predict the Other class.
", "rawMarkdown": "Thanks for your valuable comments!\nAs a biochemistry major myself, the choice of dataset is a bit strange in terms of practicality.\n\nI wonder if the second half of this competition will be a battle to see how well we can predict the Other class.", "votes": 4, "replies": [ { "id": 2519605, "postDate": "2023-11-10T07:33:55.430Z", "content": "I strongly believe top of the leaderboard will be like that.
", "rawMarkdown": "I strongly believe top of the leaderboard will be like that.", "votes": 2 } ] } ] }, { "id": 2573895, "postDate": "2023-12-25T13:04:17.497Z", "content": "the limit of this kind model 0.83(BA),maybe judge other will increase the perform
", "rawMarkdown": "the limit of this kind model 0.83(BA),maybe judge other will increase the perform", "votes": 1 }, { "id": 2518762, "postDate": "2023-11-09T14:45:12.770Z", "content": "Yes I think the same 😅
", "rawMarkdown": "Yes I think the same 😅", "votes": 1, "replies": [ { "id": 2519599, "postDate": "2023-11-10T07:25:08.947Z", "content": "Thanks for taking the time to comment.
\nMy doubts were cleared by the above posters.
In any case, it seems that we have to do our best to follow the rules of this competition.
\nLet's do our best.
", "rawMarkdown": "Thanks for taking the time to comment.\nMy doubts were cleared by the above posters.\n\nIn any case, it seems that we have to do our best to follow the rules of this competition.\n\nLet's do our best." } ] } ], "comments": [ { "id": 2519566, "author_name": "Gunes Evitan", "author_url": "", "post_date": "2023-11-10T06:58:34.097000", "content": "Based on the metric selection, Other class seems to be equally important from the clinical perspective. Even if you predict all of the samples correctly and ignore Other, you can reach 83.33% balanced accuracy. However someone with a slightly less accurate model that can capture couple Other cases might score better than you.
", "votes": 3, "replies": [ { "id": 2519597, "author_name": "Daiki Kurosu", "author_url": "", "post_date": "2023-11-10T07:23:07.600000", "content": "Thanks for your valuable comments!
\nAs a biochemistry major myself, the choice of dataset is a bit strange in terms of practicality.
I wonder if the second half of this competition will be a battle to see how well we can predict the Other class.
", "votes": 4, "replies": [ { "id": 2519605, "author_name": "Gunes Evitan", "author_url": "", "post_date": "2023-11-10T07:33:55.430000", "content": "I strongly believe top of the leaderboard will be like that.
", "votes": 2, "replies": [] } ] } ] }, { "id": 2573895, "author_name": "Toby", "author_url": "", "post_date": "2023-12-25T13:04:17.497000", "content": "the limit of this kind model 0.83(BA),maybe judge other will increase the perform
", "votes": 1, "replies": [] }, { "id": 2518762, "author_name": "Davide Camponogara", "author_url": "", "post_date": "2023-11-09T14:45:12.770000", "content": "Yes I think the same 😅
", "votes": 1, "replies": [ { "id": 2519599, "author_name": "Daiki Kurosu", "author_url": "", "post_date": "2023-11-10T07:25:08.947000", "content": "Thanks for taking the time to comment.
\nMy doubts were cleared by the above posters.
In any case, it seems that we have to do our best to follow the rules of this competition.
\nLet's do our best.
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "2518041": "What and how many are included in the label considered The Other?\n\nIn my opinion, the dataset seems to cover the four phenotypes commonly seen in ovarian cancer (although there are two for serous carcinoma) as intended.\n\nIf the Test data is sampled normally, ignoring The Other label would produce a high enough score, but I don't think that would be consistent with the questioner's intention.\n\nI would like to hear the opinions of all participants.", "2519566": "Based on the metric selection, Other class seems to be equally important from the clinical perspective. Even if you predict all of the samples correctly and ignore Other, you can reach 83.33% balanced accuracy. However someone with a slightly less accurate model that can capture couple Other cases might score better than you. ", "2573895": "the limit of this kind model 0.83(BA),maybe judge other will increase the perform", "2518762": "Yes I think the same 😅" } }