{
  "id": 232102,
  "title": "17th Place Solution",
  "content": "<p>Hi all,</p>\n<p>thanks to the organizers for this interesting and challenging competition and congratulations to the winners!</p>\n<p>Unfortunately, I dropped out of the top 10 on the way from public to private LB, but I thought that my solution might nevertheless be interesting.</p>\n<p>My approach is available in a single notebook and can be found here:<br>\n<a href=\"https://www.kaggle.com/docxian/actuarial-loss-prediction-multi-model-w2v\" target=\"_blank\">https://www.kaggle.com/docxian/actuarial-loss-prediction-multi-model-w2v</a></p>\n<p>It is basically a (equally weighted) blend of XGBoost with Gamma Loss, GBM with Gamma Loss, Random Forest and as final component a stacked ensemble of (standard) GBM and DeepLearning.</p>\n<p>The text column was tackled using a combined approach of keywords (selected by impact on target and a little bit of gut feeling) and a Word2Vec Embedding (which is calculated on the fly in the notebook =&gt; reproducibility is unfortunately not given).</p>\n<p>Thanks again and have fun!</p>",
  "comments": [
    {
      "id": 1272281,
      "author_name": "Alif Rahman",
      "author_url": "",
      "post_date": "2021-04-13T11:32:04.717000",
      "content": "<p>Thanks for sharing <a href=\"https://www.kaggle.com/docxian\" target=\"_blank\">@docxian</a> </p>",
      "votes": 0,
      "replies": []
    }
  ],
  "comment_count": 1,
  "messages": [
    {
      "id": 1270951,
      "postDate": "2021-04-12T07:21:45.763Z",
      "content": "<p>Hi all,</p>\n<p>thanks to the organizers for this interesting and challenging competition and congratulations to the winners!</p>\n<p>Unfortunately, I dropped out of the top 10 on the way from public to private LB, but I thought that my solution might nevertheless be interesting.</p>\n<p>My approach is available in a single notebook and can be found here:<br>\n<a href=\"https://www.kaggle.com/docxian/actuarial-loss-prediction-multi-model-w2v\" target=\"_blank\">https://www.kaggle.com/docxian/actuarial-loss-prediction-multi-model-w2v</a></p>\n<p>It is basically a (equally weighted) blend of XGBoost with Gamma Loss, GBM with Gamma Loss, Random Forest and as final component a stacked ensemble of (standard) GBM and DeepLearning.</p>\n<p>The text column was tackled using a combined approach of keywords (selected by impact on target and a little bit of gut feeling) and a Word2Vec Embedding (which is calculated on the fly in the notebook =&gt; reproducibility is unfortunately not given).</p>\n<p>Thanks again and have fun!</p>",
      "rawMarkdown": "Hi all,\n\nthanks to the organizers for this interesting and challenging competition and congratulations to the winners!\n\nUnfortunately, I dropped out of the top 10 on the way from public to private LB, but I thought that my solution might nevertheless be interesting.\n\nMy approach is available in a single notebook and can be found here:\n[https://www.kaggle.com/docxian/actuarial-loss-prediction-multi-model-w2v](https://www.kaggle.com/docxian/actuarial-loss-prediction-multi-model-w2v)\n\nIt is basically a (equally weighted) blend of XGBoost with Gamma Loss, GBM with Gamma Loss, Random Forest and as final component a stacked ensemble of (standard) GBM and DeepLearning.\n\nThe text column was tackled using a combined approach of keywords (selected by impact on target and a little bit of gut feeling) and a Word2Vec Embedding (which is calculated on the fly in the notebook => reproducibility is unfortunately not given).\n\nThanks again and have fun!\n",
      "votes": 8
    },
    {
      "id": 1272281,
      "postDate": "2021-04-13T11:32:04.717Z",
      "content": "<p>Thanks for sharing <a href=\"https://www.kaggle.com/docxian\" target=\"_blank\">@docxian</a> </p>",
      "rawMarkdown": "Thanks for sharing @docxian "
    }
  ],
  "complete": true,
  "source": "live",
  "fetched_at": "2026-10-19T19:14:28Z",
  "competition_id": 24764,
  "competition_slug": "actuarial-loss-estimation",
  "team_id": 6384765,
  "solution_url": "/competitions/actuarial-loss-estimation/writeups/chris-x-17th-place-solution"
}