{
  "id": 499365,
  "title": "competition_metric",
  "url": "/competitions/birdclef-2024/discussion/499365",
  "author_name": "",
  "post_date": "2024-05-01T14:06:07.153565500Z",
  "votes": 1,
  "comment_count": 4,
  "views": 0,
  "content": "<p>is there normal implementation of competition metric in sklearn?</p>",
  "messages": [
    {
      "id": "2786931",
      "postDate": "05/01/2024 14:06:07",
      "content": "<p>is there normal implementation of competition metric in sklearn?</p>",
      "rawMarkdown": "is there normal implementation of competition metric in sklearn?",
      "votes": null
    },
    {
      "id": "2788453",
      "postDate": "05/02/2024 08:43:40",
      "content": "<p>Here is the metric used by the competition: <a href=\"https://www.kaggle.com/code/metric/birdclef-roc-auc\" target=\"_blank\">ROC AUC</a>. As you can see it uses the sklearn rocauc metric which you can find here: <a href=\"https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_auc_score.html\" target=\"_blank\">Sklearn metric</a>.</p>\n<p>Happy to help!</p>",
      "rawMarkdown": "Here is the metric used by the competition: [ROC AUC](https://www.kaggle.com/code/metric/birdclef-roc-auc). As you can see it uses the sklearn rocauc metric which you can find here: [Sklearn metric](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_auc_score.html).\n\nHappy to help!",
      "votes": null
    },
    {
      "id": "2788617",
      "postDate": "05/02/2024 10:09:54",
      "content": "<p>maybe it's wrong, but this metric doesn't work with probabilities, only with indicators and labels. when trying to give it probabilities it prints 'continuous-multioutput format is not supported'</p>",
      "rawMarkdown": "maybe it's wrong, but this metric doesn't work with probabilities, only with indicators and labels. when trying to give it probabilities it prints 'continuous-multioutput format is not supported'",
      "votes": null
    },
    {
      "id": "2788628",
      "postDate": "05/02/2024 10:15:34",
      "content": "<p>I believe your true labels can only be 0 / 1, but your predictions can be probabilities!</p>",
      "rawMarkdown": "I believe your true labels can only be 0 / 1, but your predictions can be probabilities!",
      "votes": null
    },
    {
      "id": "2788995",
      "postDate": "05/02/2024 13:59:07",
      "content": "<p>thanks for answer</p>",
      "rawMarkdown": "thanks for answer",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2788453,
      "author_name": "hugodeheer",
      "author_url": "",
      "post_date": "05/02/2024 08:43:40",
      "content": "<p>Here is the metric used by the competition: <a href=\"https://www.kaggle.com/code/metric/birdclef-roc-auc\" target=\"_blank\">ROC AUC</a>. As you can see it uses the sklearn rocauc metric which you can find here: <a href=\"https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_auc_score.html\" target=\"_blank\">Sklearn metric</a>.</p>\n<p>Happy to help!</p>",
      "votes": null,
      "replies": [
        {
          "id": 2788617,
          "author_name": "korff16",
          "author_url": "",
          "post_date": "05/02/2024 10:09:54",
          "content": "<p>maybe it's wrong, but this metric doesn't work with probabilities, only with indicators and labels. when trying to give it probabilities it prints 'continuous-multioutput format is not supported'</p>",
          "votes": null,
          "replies": [
            {
              "id": 2788628,
              "author_name": "hugodeheer",
              "author_url": "",
              "post_date": "05/02/2024 10:15:34",
              "content": "<p>I believe your true labels can only be 0 / 1, but your predictions can be probabilities!</p>",
              "votes": null,
              "replies": [
                {
                  "id": 2788995,
                  "author_name": "korff16",
                  "author_url": "",
                  "post_date": "05/02/2024 13:59:07",
                  "content": "<p>thanks for answer</p>",
                  "votes": null,
                  "replies": []
                }
              ]
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2786931": "is there normal implementation of competition metric in sklearn?",
    "2788453": "Here is the metric used by the competition: [ROC AUC](https://www.kaggle.com/code/metric/birdclef-roc-auc). As you can see it uses the sklearn rocauc metric which you can find here: [Sklearn metric](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_auc_score.html).\n\nHappy to help!",
    "2788617": "maybe it's wrong, but this metric doesn't work with probabilities, only with indicators and labels. when trying to give it probabilities it prints 'continuous-multioutput format is not supported'",
    "2788628": "I believe your true labels can only be 0 / 1, but your predictions can be probabilities!",
    "2788995": "thanks for answer"
  },
  "source": "meta"
}