{
  "id": 205371,
  "title": "Choosing ensemble techniques!",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/205371",
  "author_name": "Dr. Amritpal Singh",
  "post_date": "2020-12-19T21:10:38.222000",
  "votes": 1,
  "comment_count": 0,
  "views": 0,
  "content": "<p>There are several discussion forms talking about the importance and types of ensembles, and how they are key to win the competition.<br>\nCan someone discuss how to find parameters to weigh indiviual predictions of the model?</p>\n<p>for example, in a <a href=\"https://www.kaggle.com/ilosvigil/simple-ensemble-public-private-lb-0-9658-0-9369/\" target=\"_blank\">notebook</a><br>\ndf_submission1['target'] = 0.41 * df_1['target'] + 0.04 * df_2c['target'] + 0.55 * df_3b['target']</p>\n<p>How did the author arrive at 0.41 and 0.04 values? ( Commenting on the  is not allowed, hence asking here)</p>\n<p>Another interesting discussion by <a href=\"https://www.kaggle.com/luigisaetta\" target=\"_blank\">@luigisaetta</a> <a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/166399\" target=\"_blank\">link</a> was good, but needed more light specifically on the above topic!</p>",
  "messages": [
    {
      "id": 1119262,
      "postDate": "2020-12-19T21:10:38.223Z",
      "content": "<p>There are several discussion forms talking about the importance and types of ensembles, and how they are key to win the competition.<br>\nCan someone discuss how to find parameters to weigh indiviual predictions of the model?</p>\n<p>for example, in a <a href=\"https://www.kaggle.com/ilosvigil/simple-ensemble-public-private-lb-0-9658-0-9369/\" target=\"_blank\">notebook</a><br>\ndf_submission1['target'] = 0.41 * df_1['target'] + 0.04 * df_2c['target'] + 0.55 * df_3b['target']</p>\n<p>How did the author arrive at 0.41 and 0.04 values? ( Commenting on the  is not allowed, hence asking here)</p>\n<p>Another interesting discussion by <a href=\"https://www.kaggle.com/luigisaetta\" target=\"_blank\">@luigisaetta</a> <a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/166399\" target=\"_blank\">link</a> was good, but needed more light specifically on the above topic!</p>",
      "rawMarkdown": "There are several discussion forms talking about the importance and types of ensembles, and how they are key to win the competition.\nCan someone discuss how to find parameters to weigh indiviual predictions of the model?\n\nfor example, in a [notebook](https://www.kaggle.com/ilosvigil/simple-ensemble-public-private-lb-0-9658-0-9369/)\ndf_submission1['target'] = 0.41 * df_1['target'] + 0.04 * df_2c['target'] + 0.55 * df_3b['target']\n\nHow did the author arrive at 0.41 and 0.04 values? ( Commenting on the  is not allowed, hence asking here)\n\nAnother interesting discussion by @luigisaetta [link](https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/166399) was good, but needed more light specifically on the above topic!",
      "votes": 1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1119262": "There are several discussion forms talking about the importance and types of ensembles, and how they are key to win the competition.\nCan someone discuss how to find parameters to weigh indiviual predictions of the model?\n\nfor example, in a [notebook](https://www.kaggle.com/ilosvigil/simple-ensemble-public-private-lb-0-9658-0-9369/)\ndf_submission1['target'] = 0.41 * df_1['target'] + 0.04 * df_2c['target'] + 0.55 * df_3b['target']\n\nHow did the author arrive at 0.41 and 0.04 values? ( Commenting on the  is not allowed, hence asking here)\n\nAnother interesting discussion by @luigisaetta [link](https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/166399) was good, but needed more light specifically on the above topic!"
  }
}