{
  "id": 336511,
  "title": "Loss Function Weights",
  "url": "/competitions/mayo-clinic-strip-ai/discussion/336511",
  "author_name": "moth",
  "post_date": "2022-07-11T14:59:09.360000",
  "votes": 6,
  "comment_count": 1,
  "views": 0,
  "content": "<p>As stated by the competition's hosts  we will be evaluated using a <strong>weighted multi-class logarithmic loss</strong> given by formula: </p>\n<p>$$ Log:Loss = - \\left( \\frac{\\sum_{i=1}^{M} w_{i} . \\sum_{j=1}^{N_{i}} \\frac{y_{ij}}{N_{i}}.\\ln p_{ij}}{\\sum_{i=1}^{M} w_{i}} \\right) $$</p>\n<p>Do we know if the weights <code>w_i</code> are 0.5 for each class or are they different?</p>",
  "messages": [
    {
      "id": 1851819,
      "postDate": "2022-07-11T14:59:09.360Z",
      "content": "<p>As stated by the competition's hosts  we will be evaluated using a <strong>weighted multi-class logarithmic loss</strong> given by formula: </p>\n<p>$$ Log:Loss = - \\left( \\frac{\\sum_{i=1}^{M} w_{i} . \\sum_{j=1}^{N_{i}} \\frac{y_{ij}}{N_{i}}.\\ln p_{ij}}{\\sum_{i=1}^{M} w_{i}} \\right) $$</p>\n<p>Do we know if the weights <code>w_i</code> are 0.5 for each class or are they different?</p>",
      "rawMarkdown": "As stated by the competition's hosts  we will be evaluated using a **weighted multi-class logarithmic loss** given by formula: \n\n$$ Log\\:Loss = - \\left( \\frac{\\sum_{i=1}^{M} w_{i} . \\sum_{j=1}^{N_{i}} \\frac{y_{ij}}{N_{i}}.\\ln p_{ij}}{\\sum_{i=1}^{M} w_{i}} \\right) $$\n\nDo we know if the weights `w_i` are 0.5 for each class or are they different?",
      "votes": 6
    },
    {
      "id": 1851834,
      "postDate": "2022-07-11T15:09:02.170Z",
      "content": "<p>w1=w2, see <a href=\"https://www.kaggle.com/competitions/mayo-clinic-strip-ai/discussion/335787#1848292\" target=\"_blank\">https://www.kaggle.com/competitions/mayo-clinic-strip-ai/discussion/335787#1848292</a></p>",
      "rawMarkdown": "w1=w2, see https://www.kaggle.com/competitions/mayo-clinic-strip-ai/discussion/335787#1848292",
      "votes": 3
    }
  ],
  "comments": [
    {
      "id": 1851834,
      "author_name": "Leon",
      "author_url": "",
      "post_date": "2022-07-11T15:09:02.170000",
      "content": "<p>w1=w2, see <a href=\"https://www.kaggle.com/competitions/mayo-clinic-strip-ai/discussion/335787#1848292\" target=\"_blank\">https://www.kaggle.com/competitions/mayo-clinic-strip-ai/discussion/335787#1848292</a></p>",
      "votes": 3,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1851819": "As stated by the competition's hosts  we will be evaluated using a **weighted multi-class logarithmic loss** given by formula: \n\n$$ Log\\:Loss = - \\left( \\frac{\\sum_{i=1}^{M} w_{i} . \\sum_{j=1}^{N_{i}} \\frac{y_{ij}}{N_{i}}.\\ln p_{ij}}{\\sum_{i=1}^{M} w_{i}} \\right) $$\n\nDo we know if the weights `w_i` are 0.5 for each class or are they different?",
    "1851834": "w1=w2, see https://www.kaggle.com/competitions/mayo-clinic-strip-ai/discussion/335787#1848292"
  }
}