{ "id": 167465, "title": "About my ensembles", "url": "/competitions/siim-isic-melanoma-classification/discussion/167465", "author_name": "QuantScientist", "post_date": "2020-07-16T15:35:05.817000", "votes": 29, "comment_count": 4, "views": 0, "content": "

Dear all,\nI did not train anything for this competition, but I did participate in the ISIC 2018 melanoma competition where I ranked 22nd. I shared some of my ensembling methods which I used back then. Hope this will help you in winning the competition.

\n\n

https://www.kaggle.com/solomonk/minmax-ensemble-0-9526-lb\nhttps://www.kaggle.com/solomonk/minmax-ensemble-0-9526-lb?scriptVersionId=38889837 \n(best version)

\n\n

\"\"

", "messages": [ { "id": 931969, "postDate": "2020-07-16T15:35:05.817Z", "content": "

Dear all,\nI did not train anything for this competition, but I did participate in the ISIC 2018 melanoma competition where I ranked 22nd. I shared some of my ensembling methods which I used back then. Hope this will help you in winning the competition.

\n\n

https://www.kaggle.com/solomonk/minmax-ensemble-0-9526-lb\nhttps://www.kaggle.com/solomonk/minmax-ensemble-0-9526-lb?scriptVersionId=38889837 \n(best version)

\n\n

\"\"

", "rawMarkdown": "Dear all,\nI did not train anything for this competition, but I did participate in the ISIC 2018 melanoma competition where I ranked 22nd. I shared some of my ensembling methods which I used back then. Hope this will help you in winning the competition. \n\nhttps://www.kaggle.com/solomonk/minmax-ensemble-0-9526-lb\nhttps://www.kaggle.com/solomonk/minmax-ensemble-0-9526-lb?scriptVersionId=38889837 \n(best version)\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F654629%2F4f43800e251992f5147a9633bb01456c%2F1_16bOcPi3Cq0erLcwkG6mWw.png?generation=1594913681875447&alt=media)\n\n\n", "votes": 28 }, { "id": 1119253, "postDate": "2020-12-19T21:02:42.140Z", "content": "

Hey, i came across an ensemble notebook solution[link]. It had a lot of ensmbles, that lead to good score.

\n

Code was as following -

\n
f_submission2['target'] = 0.075 * df_1['target'] + 0.1 * df_3b['target'] + 0.375 * df_4['target'] + 0.45 * df_5['target'] \n
\n

I was curious, how did you find the values for weighting indiviual component of the ensemble? How did you figure out the value 0.075, 0.375, and 0.45 ???

", "rawMarkdown": "Hey, i came across an ensemble notebook solution[[link]](https://www.kaggle.com/ilosvigil/simple-ensemble-public-private-lb-0-9658-0-9369/). It had a lot of ensmbles, that lead to good score.\n\nCode was as following - \n\n```\nf_submission2['target'] = 0.075 * df_1['target'] + 0.1 * df_3b['target'] + 0.375 * df_4['target'] + 0.45 * df_5['target'] \n\n```\nI was curious, how did you find the values for weighting indiviual component of the ensemble? How did you figure out the value 0.075, 0.375, and 0.45 ??? " }, { "id": 938696, "postDate": "2020-07-21T17:26:49.293Z", "content": "

Hey @solomonk , could you provide a link for your paper?

", "rawMarkdown": "Hey @solomonk , could you provide a link for your paper?" }, { "id": 931974, "postDate": "2020-07-16T15:43:30.317Z", "rawMarkdown": "", "votes": -1, "isDeleted": true }, { "id": 938535, "postDate": "2020-07-21T15:38:55.833Z", "content": "

thanks for sharing🙏

", "rawMarkdown": "thanks for sharing🙏 " } ], "comments": [ { "id": 1119253, "author_name": "Dr. Amritpal Singh", "author_url": "", "post_date": "2020-12-19T21:02:42.140000", "content": "

Hey, i came across an ensemble notebook solution[link]. It had a lot of ensmbles, that lead to good score.

\n

Code was as following -

\n
f_submission2['target'] = 0.075 * df_1['target'] + 0.1 * df_3b['target'] + 0.375 * df_4['target'] + 0.45 * df_5['target'] \n
\n

I was curious, how did you find the values for weighting indiviual component of the ensemble? How did you figure out the value 0.075, 0.375, and 0.45 ???

", "votes": 0, "replies": [] }, { "id": 938696, "author_name": "DimitreOliveira", "author_url": "", "post_date": "2020-07-21T17:26:49.293000", "content": "

Hey @solomonk , could you provide a link for your paper?

", "votes": 0, "replies": [] }, { "id": 931974, "author_name": "", "author_url": "", "post_date": "2020-07-16T15:43:30.317000", "content": "", "votes": -1, "replies": [] }, { "id": 938535, "author_name": "leilaSoleymani", "author_url": "", "post_date": "2020-07-21T15:38:55.833000", "content": "

thanks for sharing🙏

", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "931969": "Dear all,\nI did not train anything for this competition, but I did participate in the ISIC 2018 melanoma competition where I ranked 22nd. I shared some of my ensembling methods which I used back then. Hope this will help you in winning the competition. \n\nhttps://www.kaggle.com/solomonk/minmax-ensemble-0-9526-lb\nhttps://www.kaggle.com/solomonk/minmax-ensemble-0-9526-lb?scriptVersionId=38889837 \n(best version)\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F654629%2F4f43800e251992f5147a9633bb01456c%2F1_16bOcPi3Cq0erLcwkG6mWw.png?generation=1594913681875447&alt=media)\n\n\n", "1119253": "Hey, i came across an ensemble notebook solution[[link]](https://www.kaggle.com/ilosvigil/simple-ensemble-public-private-lb-0-9658-0-9369/). It had a lot of ensmbles, that lead to good score.\n\nCode was as following - \n\n```\nf_submission2['target'] = 0.075 * df_1['target'] + 0.1 * df_3b['target'] + 0.375 * df_4['target'] + 0.45 * df_5['target'] \n\n```\nI was curious, how did you find the values for weighting indiviual component of the ensemble? How did you figure out the value 0.075, 0.375, and 0.45 ??? ", "938696": "Hey @solomonk , could you provide a link for your paper?", "931974": "", "938535": "thanks for sharing🙏 " } }