{ "id": 54429, "title": "newbie winning strategy", "url": "/competitions/data-science-bowl-2018/discussion/54429", "author_name": "", "post_date": "2018-04-13T05:58:41.964630300Z", "votes": null, "comment_count": 7, "views": 0, "content": "

taking Heng's & Kamil's solutions as examples\nand reading comments after stage2 files were released\n i would submit a bunch of results every day\nand take a leap of faith letting the system pick the highest one\nthank you for having fun and lot's of learning stuff

", "messages": [ { "id": "313338", "postDate": "04/13/2018 05:58:41", "content": "

taking Heng's & Kamil's solutions as examples\nand reading comments after stage2 files were released\n i would submit a bunch of results every day\nand take a leap of faith letting the system pick the highest one\nthank you for having fun and lot's of learning stuff

", "rawMarkdown": "taking Heng's & Kamil's solutions as examples\nand reading comments after stage2 files were released\n i would submit a bunch of results every day\nand take a leap of faith letting the system pick the highest one\nthank you for having fun and lot's of learning stuff", "votes": null }, { "id": "313527", "postDate": "04/13/2018 12:18:55", "content": "

hahaha

", "rawMarkdown": "hahaha", "votes": null }, { "id": "314000", "postDate": "04/14/2018 10:22:57", "content": "
\n

Heng CherKeng\n 2 days ago\n Reply\n @Allen \"Heng mutters to himself “I’m going to need a bigger ensemble\" , you predicted it when i am preparing for my next post.
\n CPMP\n On Stage 2\n posted in 2018 Data Science Bowl 2 days ago\n silver\n CPMP\n 18 hours ago\n Reply\n A side effect is also that you cannot select your final 2 submissions based on LB score. You'll have to trust your local validation. I guess this is a whole new territory for some of the participants ;)<

\n
", "rawMarkdown": ">Heng CherKeng\n2 days ago\nReply\n@Allen \"Heng mutters to himself “I’m going to need a bigger ensemble\" , you predicted it when i am preparing for my next post. \nCPMP\nOn Stage 2\nposted in 2018 Data Science Bowl 2 days ago\nsilver\nCPMP\n18 hours ago\nReply\nA side effect is also that you cannot select your final 2 submissions based on LB score. You'll have to trust your local validation. I guess this is a whole new territory for some of the participants ;)<", "votes": null }, { "id": "314007", "postDate": "04/14/2018 10:57:19", "content": "

LOL.

\n\n

Sorry to ruin your dream but the system will pick your first two submissions here, see https://www.kaggle.com/c/data-science-bowl-2018/discussion/54470

", "rawMarkdown": "LOL.\n\nSorry to ruin your dream but the system will pick your first two submissions here, see https://www.kaggle.com/c/data-science-bowl-2018/discussion/54470", "votes": null }, { "id": "314026", "postDate": "04/14/2018 12:46:46", "content": "
\n

Wendy Kan\n 17 hours ago\n Reply\n In order to prevent probing, all the images in stage 2 test set are either in private leaderboard or ignored. Therefore your public leaderboard scores will always be 0.\n No. Final ranking is determined by your private leaderboard ranking.\n An empty string will work
\n It's the same as other competitions, where the private leaderboard evaluates on your two selected submissions (My Submissions-> Choose \"Use for final score\"). We'll have the system auto-select for you if you don't select.<\n Last sentence here seams to contradict.\n your post makes me smile :) thanks

\n
", "rawMarkdown": ">Wendy Kan\n17 hours ago\nReply\nIn order to prevent probing, all the images in stage 2 test set are either in private leaderboard or ignored. Therefore your public leaderboard scores will always be 0.\nNo. Final ranking is determined by your private leaderboard ranking.\nAn empty string will work \nIt's the same as other competitions, where the private leaderboard evaluates on your two selected submissions (My Submissions-> Choose \"Use for final score\"). We'll have the system auto-select for you if you don't select.<\nLast sentence here seams to contradict.\nyour post makes me smile :) thanks", "votes": null }, { "id": "314033", "postDate": "04/14/2018 13:11:07", "content": "

being and old style programmer my intuition tells me that the scoring software will do the automatic job\nbut for people who invested a lot time and effort in the competition it would be good to have the organizers write precise answer how the scoring soft is working

", "rawMarkdown": "being and old style programmer my intuition tells me that the scoring software will do the automatic job\nbut for people who invested a lot time and effort in the competition it would be good to have the organizers write precise answer how the scoring soft is working", "votes": null }, { "id": "314163", "postDate": "04/14/2018 18:48:40", "content": "

You are right

\n\n
\n

Maggie\n 3 hours ago\n Reply\n You should make your own selection of your two models, our auto-select addresses ties and will select the earliest submission.<\n all scores are ties 0's system says let's use fifo

\n
", "rawMarkdown": "You are right\n>Maggie\n3 hours ago\nReply\nYou should make your own selection of your two models, our auto-select addresses ties and will select the earliest submission.<\nall scores are ties 0's system says let's use fifo", "votes": null }, { "id": "314171", "postDate": "04/14/2018 19:24:16", "content": "

changing strategy to less is more not my idea just reading the posts

", "rawMarkdown": "changing strategy to less is more not my idea just reading the posts", "votes": null } ], "comments": [ { "id": 313527, "author_name": "wayne999", "author_url": "", "post_date": "04/13/2018 12:18:55", "content": "

hahaha

", "votes": null, "replies": [] }, { "id": 314000, "author_name": "andre1233", "author_url": "", "post_date": "04/14/2018 10:22:57", "content": "
\n

Heng CherKeng\n 2 days ago\n Reply\n @Allen \"Heng mutters to himself “I’m going to need a bigger ensemble\" , you predicted it when i am preparing for my next post.
\n CPMP\n On Stage 2\n posted in 2018 Data Science Bowl 2 days ago\n silver\n CPMP\n 18 hours ago\n Reply\n A side effect is also that you cannot select your final 2 submissions based on LB score. You'll have to trust your local validation. I guess this is a whole new territory for some of the participants ;)<

\n
", "votes": null, "replies": [] }, { "id": 314007, "author_name": "cpmpml", "author_url": "", "post_date": "04/14/2018 10:57:19", "content": "

LOL.

\n\n

Sorry to ruin your dream but the system will pick your first two submissions here, see https://www.kaggle.com/c/data-science-bowl-2018/discussion/54470

", "votes": null, "replies": [] }, { "id": 314026, "author_name": "andre1233", "author_url": "", "post_date": "04/14/2018 12:46:46", "content": "
\n

Wendy Kan\n 17 hours ago\n Reply\n In order to prevent probing, all the images in stage 2 test set are either in private leaderboard or ignored. Therefore your public leaderboard scores will always be 0.\n No. Final ranking is determined by your private leaderboard ranking.\n An empty string will work
\n It's the same as other competitions, where the private leaderboard evaluates on your two selected submissions (My Submissions-> Choose \"Use for final score\"). We'll have the system auto-select for you if you don't select.<\n Last sentence here seams to contradict.\n your post makes me smile :) thanks

\n
", "votes": null, "replies": [] }, { "id": 314033, "author_name": "andre1233", "author_url": "", "post_date": "04/14/2018 13:11:07", "content": "

being and old style programmer my intuition tells me that the scoring software will do the automatic job\nbut for people who invested a lot time and effort in the competition it would be good to have the organizers write precise answer how the scoring soft is working

", "votes": null, "replies": [] }, { "id": 314163, "author_name": "andre1233", "author_url": "", "post_date": "04/14/2018 18:48:40", "content": "

You are right

\n\n
\n

Maggie\n 3 hours ago\n Reply\n You should make your own selection of your two models, our auto-select addresses ties and will select the earliest submission.<\n all scores are ties 0's system says let's use fifo

\n
", "votes": null, "replies": [] }, { "id": 314171, "author_name": "andre1233", "author_url": "", "post_date": "04/14/2018 19:24:16", "content": "

changing strategy to less is more not my idea just reading the posts

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "313338": "taking Heng's & Kamil's solutions as examples\nand reading comments after stage2 files were released\n i would submit a bunch of results every day\nand take a leap of faith letting the system pick the highest one\nthank you for having fun and lot's of learning stuff", "313527": "hahaha", "314000": ">Heng CherKeng\n2 days ago\nReply\n@Allen \"Heng mutters to himself “I’m going to need a bigger ensemble\" , you predicted it when i am preparing for my next post. \nCPMP\nOn Stage 2\nposted in 2018 Data Science Bowl 2 days ago\nsilver\nCPMP\n18 hours ago\nReply\nA side effect is also that you cannot select your final 2 submissions based on LB score. You'll have to trust your local validation. I guess this is a whole new territory for some of the participants ;)<", "314007": "LOL.\n\nSorry to ruin your dream but the system will pick your first two submissions here, see https://www.kaggle.com/c/data-science-bowl-2018/discussion/54470", "314026": ">Wendy Kan\n17 hours ago\nReply\nIn order to prevent probing, all the images in stage 2 test set are either in private leaderboard or ignored. Therefore your public leaderboard scores will always be 0.\nNo. Final ranking is determined by your private leaderboard ranking.\nAn empty string will work \nIt's the same as other competitions, where the private leaderboard evaluates on your two selected submissions (My Submissions-> Choose \"Use for final score\"). We'll have the system auto-select for you if you don't select.<\nLast sentence here seams to contradict.\nyour post makes me smile :) thanks", "314033": "being and old style programmer my intuition tells me that the scoring software will do the automatic job\nbut for people who invested a lot time and effort in the competition it would be good to have the organizers write precise answer how the scoring soft is working", "314163": "You are right\n>Maggie\n3 hours ago\nReply\nYou should make your own selection of your two models, our auto-select addresses ties and will select the earliest submission.<\nall scores are ties 0's system says let's use fifo", "314171": "changing strategy to less is more not my idea just reading the posts" }, "source": "meta" }