![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1716193%2F01fde2cbe13fb468749a261a1d4e9b13%2FScreenshot%202023-10-07%20003421.png?generation=1696653317666412&alt=media\")
Hello,
\nI am trying to figure out which drugs are in the public test set, but couldn't find 'SB-2343' in the train data. It didn't match anything in the 'sm_name' column. I also tried 'VS-5584' which is supposed to be an alias but no match either.
\nSame for 5-Iodotubercidin, C-646, JTC-801, I tried to google the synonyms but no luck. Would anyone know what is the corresponding sm_name for these compound? Or where can I find the mapping. Thank you!
Also 'TL-HRAS-61' is also not found, 'TL_HRAS26' in 'sm_name' is a close match, are they supposed to be the same?
\nUpdate: hoho @awater1223 has the answer in his notebook, please see his answer below.
", "messages": [ { "id": "2472159", "postDate": "10/07/2023 04:38:27", "content": "Hello,
\nI am trying to figure out which drugs are in the public test set, but couldn't find 'SB-2343' in the train data. It didn't match anything in the 'sm_name' column. I also tried 'VS-5584' which is supposed to be an alias but no match either.
\nSame for 5-Iodotubercidin, C-646, JTC-801, I tried to google the synonyms but no luck. Would anyone know what is the corresponding sm_name for these compound? Or where can I find the mapping. Thank you!
Also 'TL-HRAS-61' is also not found, 'TL_HRAS26' in 'sm_name' is a close match, are they supposed to be the same?
\nUpdate: hoho @awater1223 has the answer in his notebook, please see his answer below.
", "rawMarkdown": "Hello,\n\nI am trying to figure out which drugs are in the public test set, but couldn't find 'SB-2343' in the train data. It didn't match anything in the 'sm_name' column. I also tried 'VS-5584' which is supposed to be an alias but no match either.\n\nSame for `5-Iodotubercidin`, `C-646`, `JTC-801`, I tried to google the synonyms but no luck. Would anyone know what is the corresponding sm_name for these compound? Or where can I find the mapping. Thank you!\n\n\n\n\n\nAlso 'TL-HRAS-61' is also not found, 'TL_HRAS26' in 'sm_name' is a close match, are they supposed to be the same?\n\n\n**Update: hoho @awater1223 has the answer in his notebook, please see his answer below.**", "votes": null }, { "id": "2472227", "postDate": "10/07/2023 06:17:12", "content": "@qihuaz I have tried this, you can find it in my notebook.
\nThe \"mismatching\" names were labeled, but I don't know what caused these different.
Thank you. You are awesome!
", "rawMarkdown": "Thank you. You are awesome!", "votes": null }, { "id": "2474184", "postDate": "10/09/2023 04:01:16", "content": "Sorry, how are you going to find this out? The competition description says they are random
\n\n\nPublic Test:
\n50 randomly selected compounds in B and myeloid cells
\n
\n Private Test:79 randomly selected compounds in B and myeloid cells
\n
(meaning not match those in the picture)… actually, I thought that the picture is just an example of how the data sets may look like…
", "rawMarkdown": "Sorry, how are you going to find this out? The competition description says they are random \n>Public Test:\n\n>50 randomly selected compounds in B and myeloid cells\n>Private Test:\n\n>79 randomly selected compounds in B and myeloid cells\n\n(meaning not match those in the picture)... actually, I thought that the picture is just an example of how the data sets may look like...", "votes": null }, { "id": "2474362", "postDate": "10/09/2023 06:43:35", "content": "In my understanding, they had already done the \"random selection\" and generated the figure, so it is not just an example.
\nI tried to test the data split using the LB score in my notebook v5, by masking the private set prediction to 0.
\nThe hypothesis is that the LB score should be:
mask private to zero = mask nothingmask public to zero = mask all to zeroI think the results support my opinions?
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1023301%2F785d3c1c923c005f044ad91256ae7d50%2F20231009144252.jpg?generation=1696833803831746&alt=media\")
Yes, seems to be.. Thank you a lot, its very good to know!
", "rawMarkdown": "Yes, seems to be.. Thank you a lot, its very good to know!", "votes": null }, { "id": "2479579", "postDate": "10/12/2023 17:51:12", "content": "Hi @awater1223
\nWhat do you mean by mismatching?
\nAre they using different names in train and test datasets?
Thanks
", "rawMarkdown": "Hi @awater1223 \n\nWhat do you mean by mismatching?\nAre they using different names in train and test datasets?\n\nThanks", "votes": null }, { "id": "2479989", "postDate": "10/13/2023 03:49:37", "content": "Some names in the data split figure are not match the names in the data.
\nFor example, in the figure there is a forskolin, but it should be labeled as Colforsin, which is LSM-2193 in the data.
Oh, I see.
\nthanks
", "rawMarkdown": "Oh, I see. \n\nthanks", "votes": null } ], "comments": [ { "id": 2472227, "author_name": "awater1223", "author_url": "", "post_date": "10/07/2023 06:17:12", "content": "@qihuaz I have tried this, you can find it in my notebook.
\nThe \"mismatching\" names were labeled, but I don't know what caused these different.
Thank you. You are awesome!
", "votes": null, "replies": [ { "id": 2479579, "author_name": "zhijianli", "author_url": "", "post_date": "10/12/2023 17:51:12", "content": "Hi @awater1223
\nWhat do you mean by mismatching?
\nAre they using different names in train and test datasets?
Thanks
", "votes": null, "replies": [ { "id": 2479989, "author_name": "awater1223", "author_url": "", "post_date": "10/13/2023 03:49:37", "content": "Some names in the data split figure are not match the names in the data.
\nFor example, in the figure there is a forskolin, but it should be labeled as Colforsin, which is LSM-2193 in the data.
Oh, I see.
\nthanks
", "votes": null, "replies": [] } ] } ] } ] } ] }, { "id": 2474184, "author_name": "antoninadolgorukova", "author_url": "", "post_date": "10/09/2023 04:01:16", "content": "Sorry, how are you going to find this out? The competition description says they are random
\n\n\nPublic Test:
\n50 randomly selected compounds in B and myeloid cells
\n
\n Private Test:79 randomly selected compounds in B and myeloid cells
\n
(meaning not match those in the picture)… actually, I thought that the picture is just an example of how the data sets may look like…
", "votes": null, "replies": [ { "id": 2474362, "author_name": "awater1223", "author_url": "", "post_date": "10/09/2023 06:43:35", "content": "In my understanding, they had already done the \"random selection\" and generated the figure, so it is not just an example.
\nI tried to test the data split using the LB score in my notebook v5, by masking the private set prediction to 0.
\nThe hypothesis is that the LB score should be:
mask private to zero = mask nothingmask public to zero = mask all to zeroI think the results support my opinions?
\n
Yes, seems to be.. Thank you a lot, its very good to know!
", "votes": null, "replies": [] } ] } ] } ], "raw_markdown_by_id": { "2472159": "Hello,\n\nI am trying to figure out which drugs are in the public test set, but couldn't find 'SB-2343' in the train data. It didn't match anything in the 'sm_name' column. I also tried 'VS-5584' which is supposed to be an alias but no match either.\n\nSame for `5-Iodotubercidin`, `C-646`, `JTC-801`, I tried to google the synonyms but no luck. Would anyone know what is the corresponding sm_name for these compound? Or where can I find the mapping. Thank you!\n\n\n\n\n\nAlso 'TL-HRAS-61' is also not found, 'TL_HRAS26' in 'sm_name' is a close match, are they supposed to be the same?\n\n\n**Update: hoho @awater1223 has the answer in his notebook, please see his answer below.**", "2472227": "qihuaz I have tried this, you can find it in my [notebook](https://www.kaggle.com/code/awater1223/op2-00-basic-metadata-eda#Data-split).\nThe \"mismatching\" names were labeled, but I don't know what caused these different.", "2472915": "Thank you. You are awesome!", "2474184": "Sorry, how are you going to find this out? The competition description says they are random \n>Public Test:\n\n>50 randomly selected compounds in B and myeloid cells\n>Private Test:\n\n>79 randomly selected compounds in B and myeloid cells\n\n(meaning not match those in the picture)... actually, I thought that the picture is just an example of how the data sets may look like...", "2474362": "In my understanding, they had already done the \"random selection\" and generated the figure, so it is not just an example.\n\nI tried to test the data split using the LB score in my [notebook](https://www.kaggle.com/code/awater1223/op2-00-basic-metadata-eda) v5, by masking the private set prediction to 0.\nThe hypothesis is that the LB score should be:\n- `mask private to zero` = `mask nothing`\n- `mask public to zero` = `mask all to zero`\n\nI think the results support my opinions?\n\n", "2474458": "Yes, seems to be.. Thank you a lot, its very good to know!", "2479579": "Hi @awater1223 \n\nWhat do you mean by mismatching?\nAre they using different names in train and test datasets?\n\nThanks", "2479989": "Some names in the **data split figure** are not match the names in the data.\nFor example, in the figure there is a `forskolin`, but it should be labeled as `Colforsin`, which is `LSM-2193` in the data.", "2481038": "Oh, I see. \n\nthanks" }, "source": "meta" }