{
  "id": 165623,
  "title": "Ensembling (for Vision) models",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/165623",
  "author_name": "",
  "post_date": "2020-07-10T11:41:03.180925100Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Recently I learnt about different ensembling techniques like bagging, stacking, boosting etc etc\nI was working on object detection and classification problem and applied several models.\nNow how to use these models together(ensembling?) is my doubt.\nI have read several Kaggle discussions where the author just says that they have ensembles he models to get a better accuracy model, how do they do that?\nSorry if the question sounds too basic.</p>",
  "messages": [
    {
      "id": "922881",
      "postDate": "07/10/2020 11:41:03",
      "content": "<p>Recently I learnt about different ensembling techniques like bagging, stacking, boosting etc etc\nI was working on object detection and classification problem and applied several models.\nNow how to use these models together(ensembling?) is my doubt.\nI have read several Kaggle discussions where the author just says that they have ensembles he models to get a better accuracy model, how do they do that?\nSorry if the question sounds too basic.</p>",
      "rawMarkdown": "Recently I learnt about different ensembling techniques like bagging, stacking, boosting etc etc\nI was working on object detection and classification problem and applied several models.\nNow how to use these models together(ensembling?) is my doubt.\nI have read several Kaggle discussions where the author just says that they have ensembles he models to get a better accuracy model, how do they do that?\nSorry if the question sounds too basic.",
      "votes": null
    },
    {
      "id": "922912",
      "postDate": "07/10/2020 12:12:37",
      "content": "<p>By using bagging, stacking, averaging or boosting just like you mention ;)</p>",
      "rawMarkdown": "By using bagging, stacking, averaging or boosting just like you mention ;)",
      "votes": null
    },
    {
      "id": "922918",
      "postDate": "07/10/2020 12:17:45",
      "content": "<p><a href=\"/mrinath\">@mrinath</a> Did you check public kernels such as these:\n- <a href=\"https://www.kaggle.com/khoongweihao/siim-isic-multiple-model-training-stacking#Simple-Blend-of-Predictions\">https://www.kaggle.com/khoongweihao/siim-isic-multiple-model-training-stacking#Simple-Blend-of-Predictions</a>\n- <a href=\"https://www.kaggle.com/arthurtok/introduction-to-ensembling-stacking-in-python\">https://www.kaggle.com/arthurtok/introduction-to-ensembling-stacking-in-python</a></p>",
      "rawMarkdown": "mrinath Did you check public kernels such as these:\n- https://www.kaggle.com/khoongweihao/siim-isic-multiple-model-training-stacking#Simple-Blend-of-Predictions\n- https://www.kaggle.com/arthurtok/introduction-to-ensembling-stacking-in-python",
      "votes": null
    },
    {
      "id": "923000",
      "postDate": "07/10/2020 13:24:24",
      "content": "<p>Also see my new post <a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/165653\">Ensemble using Power Averaging</a></p>",
      "rawMarkdown": "Also see my new post [Ensemble using Power Averaging](https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/165653)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 922912,
      "author_name": "group16",
      "author_url": "",
      "post_date": "07/10/2020 12:12:37",
      "content": "<p>By using bagging, stacking, averaging or boosting just like you mention ;)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 922918,
      "author_name": "sirishks",
      "author_url": "",
      "post_date": "07/10/2020 12:17:45",
      "content": "<p><a href=\"/mrinath\">@mrinath</a> Did you check public kernels such as these:\n- <a href=\"https://www.kaggle.com/khoongweihao/siim-isic-multiple-model-training-stacking#Simple-Blend-of-Predictions\">https://www.kaggle.com/khoongweihao/siim-isic-multiple-model-training-stacking#Simple-Blend-of-Predictions</a>\n- <a href=\"https://www.kaggle.com/arthurtok/introduction-to-ensembling-stacking-in-python\">https://www.kaggle.com/arthurtok/introduction-to-ensembling-stacking-in-python</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 923000,
      "author_name": "sirishks",
      "author_url": "",
      "post_date": "07/10/2020 13:24:24",
      "content": "<p>Also see my new post <a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/165653\">Ensemble using Power Averaging</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "922881": "Recently I learnt about different ensembling techniques like bagging, stacking, boosting etc etc\nI was working on object detection and classification problem and applied several models.\nNow how to use these models together(ensembling?) is my doubt.\nI have read several Kaggle discussions where the author just says that they have ensembles he models to get a better accuracy model, how do they do that?\nSorry if the question sounds too basic.",
    "922912": "By using bagging, stacking, averaging or boosting just like you mention ;)",
    "922918": "mrinath Did you check public kernels such as these:\n- https://www.kaggle.com/khoongweihao/siim-isic-multiple-model-training-stacking#Simple-Blend-of-Predictions\n- https://www.kaggle.com/arthurtok/introduction-to-ensembling-stacking-in-python",
    "923000": "Also see my new post [Ensemble using Power Averaging](https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/165653)"
  },
  "source": "meta"
}