Pretty late in the game to make any big changes. But recently it hit me that after all the time spent on the competition, the final shakeup could lay everything to waste (DUHHHH). I'm a little slow on the uptake. There have been several topics on LB shakeup and the differences in distribution.
\nI spent the time I had learning to implement new algorithms and doing feature engineering. Comes down to lack of prioritization and clear direction in my case. But it's great to learn! I look forward to reading the brilliant solutions at the end of the competition
\nI have explored some of the differences in variable distribution between the training data and private test data in the following notebook
\nhttps://www.kaggle.com/code/illidan7/amex-final-shakeup-how-screwed-are-you
\nGood luck to all!
\nCredits:
\nPretty late in the game to make any big changes. But recently it hit me that after all the time spent on the competition, the final shakeup could lay everything to waste (DUHHHH). I'm a little slow on the uptake. There have been several topics on LB shakeup and the differences in distribution.
\nI spent the time I had learning to implement new algorithms and doing feature engineering. Comes down to lack of prioritization and clear direction in my case. But it's great to learn! I look forward to reading the brilliant solutions at the end of the competition
\nI have explored some of the differences in variable distribution between the training data and private test data in the following notebook
\nhttps://www.kaggle.com/code/illidan7/amex-final-shakeup-how-screwed-are-you
\nGood luck to all!
\nCredits:
\nWhatever happens in the shakeup, the positive thing is the things you've learned. You've already identified some of those. Sometimes the shakeup itself can be the biggest learning experience, although, of course, we're all hoping that we will shake up, not down!
", "rawMarkdown": "Whatever happens in the shakeup, the positive thing is the things you've learned. You've already identified some of those. Sometimes the shakeup itself can be the biggest learning experience, although, of course, we're all hoping that we will shake up, not down!", "votes": null }, { "id": "1904884", "postDate": "08/18/2022 15:10:11", "content": "Yeah definitely trying to focus more on the learning. And very grateful to the Kaggle community for that!
\nIt will still sting the competitive side of me though 😄
", "rawMarkdown": "Yeah definitely trying to focus more on the learning. And very grateful to the Kaggle community for that!\n\nIt will still sting the competitive side of me though 😄", "votes": null } ], "comments": [ { "id": 1904601, "author_name": "datahobbit", "author_url": "", "post_date": "08/18/2022 10:31:43", "content": "Whatever happens in the shakeup, the positive thing is the things you've learned. You've already identified some of those. Sometimes the shakeup itself can be the biggest learning experience, although, of course, we're all hoping that we will shake up, not down!
", "votes": null, "replies": [ { "id": 1904884, "author_name": "illidan7", "author_url": "", "post_date": "08/18/2022 15:10:11", "content": "Yeah definitely trying to focus more on the learning. And very grateful to the Kaggle community for that!
\nIt will still sting the competitive side of me though 😄
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1903663": "Pretty late in the game to make any big changes. But recently it hit me that after all the time spent on the competition, the final shakeup could lay everything to waste (DUHHHH). I'm a little slow on the uptake. There have been several topics on LB shakeup and the differences in distribution.\n\nI spent the time I had learning to implement new algorithms and doing feature engineering. Comes down to lack of prioritization and clear direction in my case. But it's great to learn! I look forward to reading the brilliant solutions at the end of the competition\n\nI have explored some of the differences in variable distribution between the training data and private test data in the following notebook\n\nhttps://www.kaggle.com/code/illidan7/amex-final-shakeup-how-screwed-are-you\n\nGood luck to all! \n____________________\nCredits:\n\n- https://www.kaggle.com/code/zakopur0/adversarial-validation-private-vs-public\n- https://www.kaggle.com/datasets/raddar/amex-data-integer-dtypes-parquet-format", "1904601": "Whatever happens in the shakeup, the positive thing is the things you've learned. You've already identified some of those. Sometimes the shakeup itself can be the biggest learning experience, although, of course, we're all hoping that we will shake up, not down!", "1904884": "Yeah definitely trying to focus more on the learning. And very grateful to the Kaggle community for that!\n\nIt will still sting the competitive side of me though 😄" }, "source": "meta" }