CREDIT: goes to @jzhllin based on his amazing post-processing sorting strategy: https://www.kaggle.com/competitions/home-credit-credit-risk-model-stability/discussion/507959
\nNOTEBOOK: https://www.kaggle.com/code/adegladius/unofficial-winning-solution-0-618-on-private-lb
", "messages": [ { "id": 2840585, "postDate": "2024-05-28T06:44:18.027Z", "content": "CREDIT: goes to @jzhllin based on his amazing post-processing sorting strategy: https://www.kaggle.com/competitions/home-credit-credit-risk-model-stability/discussion/507959
\nNOTEBOOK: https://www.kaggle.com/code/adegladius/unofficial-winning-solution-0-618-on-private-lb
", "rawMarkdown": "**CREDIT:** goes to @jzhllin based on his amazing post-processing sorting strategy: https://www.kaggle.com/competitions/home-credit-credit-risk-model-stability/discussion/507959\n\n**NOTEBOOK:** https://www.kaggle.com/code/adegladius/unofficial-winning-solution-0-618-on-private-lb", "votes": 1 }, { "id": 2840630, "postDate": "2024-05-28T07:11:59.953Z", "content": "You gave credit to @jzhllin for his script but you are not on the same team and I don’t see it publicly posted? So where did you first see it?
\nEdit: Its a late submission notebook only, which makes sense now.
", "rawMarkdown": "You gave credit to @jzhllin for his script but you are not on the same team and I don’t see it publicly posted? So where did you first see it?\n\nEdit: Its a late submission notebook only, which makes sense now.", "votes": 2, "replies": [ { "id": 2840683, "postDate": "2024-05-28T07:48:44.100Z", "content": "Yes,I don't have the complete open-source code because code 0.585 or 0.584 can be found in the open-source area，It is obvious that this Notebook replicates the metric hack solution I mentioned, which is to add the post-processing code I mentioned on top of the original open-source code of 0.584 or 0.585；You can take a look at the discussion content I posted, I mentioned this point
", "rawMarkdown": "Yes,I don't have the complete open-source code because code 0.585 or 0.584 can be found in the open-source area，It is obvious that this Notebook replicates the metric hack solution I mentioned, which is to add the post-processing code I mentioned on top of the original open-source code of 0.584 or 0.585；You can take a look at the discussion content I posted, I mentioned this point", "replies": [ { "id": 2841415, "postDate": "2024-05-28T15:14:53.647Z", "content": "@jzhllin Exactly, that's exactly what happened. @julianmukaj you can as well follow @jzhllin earlier discussion for clarification on the original open-source code and additional post-processing code mentioned by @jzhllin.
", "rawMarkdown": "@jzhllin Exactly, that's exactly what happened. @julianmukaj you can as well follow @jzhllin earlier discussion for clarification on the original open-source code and additional post-processing code mentioned by @jzhllin. " }, { "id": 2841445, "postDate": "2024-05-28T15:22:51.917Z", "content": "@julianmukaj Again, inside this notebook: https://www.kaggle.com/code/adegladius/unofficial-winning-solution-0-618-on-private-lb there's a link to discussion of this post-processing strategy created by @jzhllin. Hope it's clear now.
", "rawMarkdown": "@julianmukaj Again, inside this notebook: https://www.kaggle.com/code/adegladius/unofficial-winning-solution-0-618-on-private-lb there's a link to discussion of this post-processing strategy created by @jzhllin. Hope it's clear now. " }, { "id": 2841893, "postDate": "2024-05-28T18:27:13.357Z", "rawMarkdown": "", "isDeleted": true } ] } ] } ], "comments": [ { "id": 2840630, "author_name": "JM", "author_url": "", "post_date": "2024-05-28T07:11:59.953000", "content": "You gave credit to @jzhllin for his script but you are not on the same team and I don’t see it publicly posted? So where did you first see it?
\nEdit: Its a late submission notebook only, which makes sense now.
", "votes": 2, "replies": [ { "id": 2840683, "author_name": "按59", "author_url": "", "post_date": "2024-05-28T07:48:44.100000", "content": "Yes,I don't have the complete open-source code because code 0.585 or 0.584 can be found in the open-source area，It is obvious that this Notebook replicates the metric hack solution I mentioned, which is to add the post-processing code I mentioned on top of the original open-source code of 0.584 or 0.585；You can take a look at the discussion content I posted, I mentioned this point
", "votes": 0, "replies": [ { "id": 2841415, "author_name": "MICADEE", "author_url": "", "post_date": "2024-05-28T15:14:53.647000", "content": "@jzhllin Exactly, that's exactly what happened. @julianmukaj you can as well follow @jzhllin earlier discussion for clarification on the original open-source code and additional post-processing code mentioned by @jzhllin.
", "votes": 0, "replies": [] }, { "id": 2841445, "author_name": "MICADEE", "author_url": "", "post_date": "2024-05-28T15:22:51.917000", "content": "@julianmukaj Again, inside this notebook: https://www.kaggle.com/code/adegladius/unofficial-winning-solution-0-618-on-private-lb there's a link to discussion of this post-processing strategy created by @jzhllin. Hope it's clear now.
", "votes": 0, "replies": [] }, { "id": 2841893, "author_name": "", "author_url": "", "post_date": "2024-05-28T18:27:13.357000", "content": "", "votes": 0, "replies": [] } ] } ] } ], "raw_markdown_by_id": { "2840585": "**CREDIT:** goes to @jzhllin based on his amazing post-processing sorting strategy: https://www.kaggle.com/competitions/home-credit-credit-risk-model-stability/discussion/507959\n\n**NOTEBOOK:** https://www.kaggle.com/code/adegladius/unofficial-winning-solution-0-618-on-private-lb", "2840630": "You gave credit to @jzhllin for his script but you are not on the same team and I don’t see it publicly posted? So where did you first see it?\n\nEdit: Its a late submission notebook only, which makes sense now." } }