{
  "id": 114019,
  "title": "Tune LGBM parameters.",
  "url": "/competitions/ashrae-energy-prediction/discussion/114019",
  "author_name": "",
  "post_date": "2019-10-23T15:56:22.407143300Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I tuned my parameters and increased my score by 0.02. \nI used the Bayesian Optimization.</p>\n\n<p>You can get the code from \n<a href=\"https://github.com/fmfn/BayesianOptimization\">https://github.com/fmfn/BayesianOptimization</a>  </p>",
  "messages": [
    {
      "id": "655878",
      "postDate": "10/23/2019 15:56:22",
      "content": "<p>I tuned my parameters and increased my score by 0.02. \nI used the Bayesian Optimization.</p>\n\n<p>You can get the code from \n<a href=\"https://github.com/fmfn/BayesianOptimization\">https://github.com/fmfn/BayesianOptimization</a>  </p>",
      "rawMarkdown": "I tuned my parameters and increased my score by 0.02. \nI used the Bayesian Optimization.\n\nYou can get the code from \nhttps://github.com/fmfn/BayesianOptimization",
      "votes": null
    },
    {
      "id": "656386",
      "postDate": "10/24/2019 07:31:22",
      "content": "<p>Try to generate multiple outputs for different configuration of your model and then you can use GMEAN of low correlated values.</p>",
      "rawMarkdown": "Try to generate multiple outputs for different configuration of your model and then you can use GMEAN of low correlated values.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 656386,
      "author_name": "shishu1421",
      "author_url": "",
      "post_date": "10/24/2019 07:31:22",
      "content": "<p>Try to generate multiple outputs for different configuration of your model and then you can use GMEAN of low correlated values.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "655878": "I tuned my parameters and increased my score by 0.02. \nI used the Bayesian Optimization.\n\nYou can get the code from \nhttps://github.com/fmfn/BayesianOptimization",
    "656386": "Try to generate multiple outputs for different configuration of your model and then you can use GMEAN of low correlated values."
  },
  "source": "meta"
}