I spent couple of hours to understand how NA's have been generated in the dataset. And my oh my there are so much to discover here. I covered most of my work so far in the notebook:
\nhttps://www.kaggle.com/raddar/understanding-na-values-in-amex-competition
\nEnjoy!
", "messages": [ { "id": "1824707", "postDate": "06/18/2022 15:00:14", "content": "I spent couple of hours to understand how NA's have been generated in the dataset. And my oh my there are so much to discover here. I covered most of my work so far in the notebook:
\nhttps://www.kaggle.com/raddar/understanding-na-values-in-amex-competition
\nEnjoy!
", "rawMarkdown": "I spent couple of hours to understand how NA's have been generated in the dataset. And my oh my there are so much to discover here. I covered most of my work so far in the notebook:\n\nhttps://www.kaggle.com/raddar/understanding-na-values-in-amex-competition\n\nEnjoy!", "votes": null }, { "id": "1824734", "postDate": "06/18/2022 15:23:23", "content": "Thank you @raddar for all the good work you are sharing( discussions, notebooks ) , it always helps , keep it up 🙌
", "rawMarkdown": "Thank you @raddar for all the good work you are sharing( discussions, notebooks ) , it always helps , keep it up 🙌", "votes": null }, { "id": "1824736", "postDate": "06/18/2022 15:24:01", "content": "I have mixed feelings)) upvoted - did the same study
\nbut sometimes I think to not \"waste\" time in this competition during june and july and just come back in august and take everything from you notebooks and other people sharings.
mixed feelings…
", "rawMarkdown": "I have mixed feelings)) upvoted - did the same study\nbut sometimes I think to not \"waste\" time in this competition during june and july and just come back in august and take everything from you notebooks and other people sharings.\n\nmixed feelings...", "votes": null }, { "id": "1824746", "postDate": "06/18/2022 15:32:08", "content": "Well. this does happen in most of the competitions, doesn't it :) I think this does less harm than good performing public models, as you do have to figure out yourself how to use the information you get from this kind of work
", "rawMarkdown": "Well. this does happen in most of the competitions, doesn't it :) I think this does less harm than good performing public models, as you do have to figure out yourself how to use the information you get from this kind of work", "votes": null }, { "id": "1824752", "postDate": "06/18/2022 15:37:44", "content": "Interesting work! Your discovery has inspired me a lot.
", "rawMarkdown": "Interesting work! Your discovery has inspired me a lot.", "votes": null }, { "id": "1824754", "postDate": "06/18/2022 15:38:00", "content": "You are totally right to post all such findings - many people (and I'm too) learning a lot from such posts and notebooks. Please continue to do so.
", "rawMarkdown": "You are totally right to post all such findings - many people (and I'm too) learning a lot from such posts and notebooks. Please continue to do so.", "votes": null }, { "id": "1824837", "postDate": "06/18/2022 16:38:14", "content": "@raddar you are doing the most. Thanks!
", "rawMarkdown": "raddar you are doing the most. Thanks!", "votes": null }, { "id": "1824943", "postDate": "06/18/2022 19:23:36", "content": "nice finding!
", "rawMarkdown": "nice finding!", "votes": null }, { "id": "1825617", "postDate": "06/19/2022 14:46:54", "content": "Thank you for sharing @raddar
", "rawMarkdown": "Thank you for sharing @raddar", "votes": null }, { "id": "1826160", "postDate": "06/20/2022 05:17:19", "content": "Thanks for sharing.
", "rawMarkdown": "Thanks for sharing.", "votes": null }, { "id": "1827130", "postDate": "06/20/2022 21:11:20", "content": "i wonder if you use transformer to learn to predict masked input (on both train and test csv) would help to fill in missing values?
", "rawMarkdown": "i wonder if you use transformer to learn to predict masked input (on both train and test csv) would help to fill in missing values?", "votes": null }, { "id": "1827139", "postDate": "06/20/2022 21:30:49", "content": "I guess it could. most NA patterns looks straightforward to impute
", "rawMarkdown": "I guess it could. most NA patterns looks straightforward to impute", "votes": null }, { "id": "1828160", "postDate": "06/21/2022 14:12:50", "content": "Really needed that, Thanks!
", "rawMarkdown": "Really needed that, Thanks!", "votes": null }, { "id": "1828230", "postDate": "06/21/2022 15:12:24", "content": "Thanks for sharing, always great content to learn from @raddar
", "rawMarkdown": "Thanks for sharing, always great content to learn from @raddar", "votes": null }, { "id": "1849989", "postDate": "07/10/2022 01:53:26", "content": "You are a gentleman and a scholar, thanks for all the contributions to the community.
", "rawMarkdown": "You are a gentleman and a scholar, thanks for all the contributions to the community.", "votes": null } ], "comments": [ { "id": 1824734, "author_name": "mhslearner", "author_url": "", "post_date": "06/18/2022 15:23:23", "content": "Thank you @raddar for all the good work you are sharing( discussions, notebooks ) , it always helps , keep it up 🙌
", "votes": null, "replies": [] }, { "id": 1824736, "author_name": "kyakovlev", "author_url": "", "post_date": "06/18/2022 15:24:01", "content": "I have mixed feelings)) upvoted - did the same study
\nbut sometimes I think to not \"waste\" time in this competition during june and july and just come back in august and take everything from you notebooks and other people sharings.
mixed feelings…
", "votes": null, "replies": [ { "id": 1824746, "author_name": "raddar", "author_url": "", "post_date": "06/18/2022 15:32:08", "content": "Well. this does happen in most of the competitions, doesn't it :) I think this does less harm than good performing public models, as you do have to figure out yourself how to use the information you get from this kind of work
", "votes": null, "replies": [] }, { "id": 1824754, "author_name": "kyakovlev", "author_url": "", "post_date": "06/18/2022 15:38:00", "content": "You are totally right to post all such findings - many people (and I'm too) learning a lot from such posts and notebooks. Please continue to do so.
", "votes": null, "replies": [] } ] }, { "id": 1824752, "author_name": "x2c2y4", "author_url": "", "post_date": "06/18/2022 15:37:44", "content": "Interesting work! Your discovery has inspired me a lot.
", "votes": null, "replies": [] }, { "id": 1824837, "author_name": "wuuthraad", "author_url": "", "post_date": "06/18/2022 16:38:14", "content": "@raddar you are doing the most. Thanks!
", "votes": null, "replies": [] }, { "id": 1824943, "author_name": "aliabdin1", "author_url": "", "post_date": "06/18/2022 19:23:36", "content": "nice finding!
", "votes": null, "replies": [] }, { "id": 1825617, "author_name": "arti1117", "author_url": "", "post_date": "06/19/2022 14:46:54", "content": "Thank you for sharing @raddar
", "votes": null, "replies": [] }, { "id": 1826160, "author_name": "deba87", "author_url": "", "post_date": "06/20/2022 05:17:19", "content": "Thanks for sharing.
", "votes": null, "replies": [] }, { "id": 1827130, "author_name": "hengck23", "author_url": "", "post_date": "06/20/2022 21:11:20", "content": "i wonder if you use transformer to learn to predict masked input (on both train and test csv) would help to fill in missing values?
", "votes": null, "replies": [ { "id": 1827139, "author_name": "raddar", "author_url": "", "post_date": "06/20/2022 21:30:49", "content": "I guess it could. most NA patterns looks straightforward to impute
", "votes": null, "replies": [] } ] }, { "id": 1828160, "author_name": "rohit2710", "author_url": "", "post_date": "06/21/2022 14:12:50", "content": "Really needed that, Thanks!
", "votes": null, "replies": [] }, { "id": 1828230, "author_name": "lucafaccin", "author_url": "", "post_date": "06/21/2022 15:12:24", "content": "Thanks for sharing, always great content to learn from @raddar
", "votes": null, "replies": [] }, { "id": 1849989, "author_name": "scharlesworth", "author_url": "", "post_date": "07/10/2022 01:53:26", "content": "You are a gentleman and a scholar, thanks for all the contributions to the community.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1824707": "I spent couple of hours to understand how NA's have been generated in the dataset. And my oh my there are so much to discover here. I covered most of my work so far in the notebook:\n\nhttps://www.kaggle.com/raddar/understanding-na-values-in-amex-competition\n\nEnjoy!", "1824734": "Thank you @raddar for all the good work you are sharing( discussions, notebooks ) , it always helps , keep it up 🙌", "1824736": "I have mixed feelings)) upvoted - did the same study\nbut sometimes I think to not \"waste\" time in this competition during june and july and just come back in august and take everything from you notebooks and other people sharings.\n\nmixed feelings...", "1824746": "Well. this does happen in most of the competitions, doesn't it :) I think this does less harm than good performing public models, as you do have to figure out yourself how to use the information you get from this kind of work", "1824752": "Interesting work! Your discovery has inspired me a lot.", "1824754": "You are totally right to post all such findings - many people (and I'm too) learning a lot from such posts and notebooks. Please continue to do so.", "1824837": "raddar you are doing the most. Thanks!", "1824943": "nice finding!", "1825617": "Thank you for sharing @raddar", "1826160": "Thanks for sharing.", "1827130": "i wonder if you use transformer to learn to predict masked input (on both train and test csv) would help to fill in missing values?", "1827139": "I guess it could. most NA patterns looks straightforward to impute", "1828160": "Really needed that, Thanks!", "1828230": "Thanks for sharing, always great content to learn from @raddar", "1849989": "You are a gentleman and a scholar, thanks for all the contributions to the community." }, "source": "meta" }