{
  "id": 174080,
  "title": "Ensuring better score than previously submitted submission file",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/174080",
  "author_name": "",
  "post_date": "2020-08-12T08:12:33.578597Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>How can I check that my new submission file's accuracy/score is better than my previous submitted submission file? Is there any way? </p>\n<p>TIA.</p>",
  "messages": [
    {
      "id": "967387",
      "postDate": "08/12/2020 08:12:33",
      "content": "<p>How can I check that my new submission file's accuracy/score is better than my previous submitted submission file? Is there any way? </p>\n<p>TIA.</p>",
      "rawMarkdown": "How can I check that my new submission file's accuracy/score is better than my previous submitted submission file? Is there any way? \n\nTIA.",
      "votes": null
    },
    {
      "id": "967540",
      "postDate": "08/12/2020 10:18:34",
      "content": "<p>If I understand your question correctly the standard way would be to use a cross-validation procedure: </p>\n<p><a href=\"https://scikit-learn.org/stable/modules/cross_validation.html\" target=\"_blank\">https://scikit-learn.org/stable/modules/cross_validation.html</a></p>\n<p>Although this by no means guarantees that your public leader board score of a submission with a larger cross-validation score will be higher than the public leader board of a model with a lower cross-validation score.</p>",
      "rawMarkdown": "If I understand your question correctly the standard way would be to use a cross-validation procedure: \n\nhttps://scikit-learn.org/stable/modules/cross_validation.html\n\nAlthough this by no means guarantees that your public leader board score of a submission with a larger cross-validation score will be higher than the public leader board of a model with a lower cross-validation score.",
      "votes": null
    },
    {
      "id": "967728",
      "postDate": "08/12/2020 13:12:44",
      "content": "<p>You can sort your submissions by score.  Even if they display the same score they will be ranked by the actual score.</p>",
      "rawMarkdown": "You can sort your submissions by score.  Even if they display the same score they will be ranked by the actual score.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 967540,
      "author_name": "fchmiel",
      "author_url": "",
      "post_date": "08/12/2020 10:18:34",
      "content": "<p>If I understand your question correctly the standard way would be to use a cross-validation procedure: </p>\n<p><a href=\"https://scikit-learn.org/stable/modules/cross_validation.html\" target=\"_blank\">https://scikit-learn.org/stable/modules/cross_validation.html</a></p>\n<p>Although this by no means guarantees that your public leader board score of a submission with a larger cross-validation score will be higher than the public leader board of a model with a lower cross-validation score.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 967728,
      "author_name": "cpmpml",
      "author_url": "",
      "post_date": "08/12/2020 13:12:44",
      "content": "<p>You can sort your submissions by score.  Even if they display the same score they will be ranked by the actual score.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "967387": "How can I check that my new submission file's accuracy/score is better than my previous submitted submission file? Is there any way? \n\nTIA.",
    "967540": "If I understand your question correctly the standard way would be to use a cross-validation procedure: \n\nhttps://scikit-learn.org/stable/modules/cross_validation.html\n\nAlthough this by no means guarantees that your public leader board score of a submission with a larger cross-validation score will be higher than the public leader board of a model with a lower cross-validation score.",
    "967728": "You can sort your submissions by score.  Even if they display the same score they will be ranked by the actual score."
  },
  "source": "meta"
}