{
  "id": 312875,
  "title": "Question About The Test Data Labels",
  "url": "/competitions/herbarium-2022-fgvc9/discussion/312875",
  "author_name": "beasterea1109",
  "post_date": "2022-03-14T14:35:49.585000",
  "votes": 0,
  "comment_count": 4,
  "views": 0,
  "content": "<p>I wonder if the \"category_id\" we have to predict for the test data is all included in the train data we are given. <br>\nIf so, I was considering to check the dependency between the \"family\", \"genus\" and \"species\" to generate a boolean mask for sub classifiers for genus and species.</p>",
  "messages": [
    {
      "id": 1735976,
      "postDate": "2022-03-26T19:03:04.620Z",
      "content": "<p>taxonomy ranks are hierarchical, so each family have many genera, and each genus have many species.</p>\n<p>each given species will be a member of a single family and genus.</p>",
      "rawMarkdown": "taxonomy ranks are hierarchical, so each family have many genera, and each genus have many species.\n\neach given species will be a member of a single family and genus."
    },
    {
      "id": 1722597,
      "postDate": "2022-03-14T17:34:19.010Z",
      "content": "<p>Yes, they are all included in the training data. Family and genus names are all unique, but species names are dependent on genus names. (For example, there could be same species name for different genus names, such as <em>Abies    amabilis</em> vs <em>Cyperus amabilis</em>). </p>",
      "rawMarkdown": "Yes, they are all included in the training data. Family and genus names are all unique, but species names are dependent on genus names. (For example, there could be same species name for different genus names, such as *Abies\tamabilis* vs *Cyperus amabilis*). ",
      "replies": [
        {
          "id": 1750058,
          "postDate": "2022-04-09T09:22:44.590Z",
          "content": "<p>How come I saw this comment now,, <br>\nI came to a result that each of the species belongs to only one family, according to the train dataset.<br>\nIs this the same for the test dataset too??<br>\nThank You.</p>",
          "rawMarkdown": "How come I saw this comment now,, \nI came to a result that each of the species belongs to only one family, according to the train dataset.\nIs this the same for the test dataset too??\nThank You."
        },
        {
          "id": 1778946,
          "postDate": "2022-05-06T00:14:15.127Z",
          "content": "<p>Hi <a href=\"https://www.kaggle.com/beasterea1109\" target=\"_blank\">@beasterea1109</a>!<br>\nYes, that is correct! Each species belongs to only one genus and one family.</p>",
          "rawMarkdown": "Hi @beasterea1109!\nYes, that is correct! Each species belongs to only one genus and one family."
        }
      ]
    },
    {
      "id": 1722432,
      "postDate": "2022-03-14T14:35:49.587Z",
      "content": "<p>I wonder if the \"category_id\" we have to predict for the test data is all included in the train data we are given. <br>\nIf so, I was considering to check the dependency between the \"family\", \"genus\" and \"species\" to generate a boolean mask for sub classifiers for genus and species.</p>",
      "rawMarkdown": "I wonder if the \"category_id\" we have to predict for the test data is all included in the train data we are given. \nIf so, I was considering to check the dependency between the \"family\", \"genus\" and \"species\" to generate a boolean mask for sub classifiers for genus and species."
    }
  ],
  "comments": [
    {
      "id": 1735976,
      "author_name": "rvalieris",
      "author_url": "",
      "post_date": "2022-03-26T19:03:04.620000",
      "content": "<p>taxonomy ranks are hierarchical, so each family have many genera, and each genus have many species.</p>\n<p>each given species will be a member of a single family and genus.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 1722597,
      "author_name": "John Park",
      "author_url": "",
      "post_date": "2022-03-14T17:34:19.010000",
      "content": "<p>Yes, they are all included in the training data. Family and genus names are all unique, but species names are dependent on genus names. (For example, there could be same species name for different genus names, such as <em>Abies    amabilis</em> vs <em>Cyperus amabilis</em>). </p>",
      "votes": 0,
      "replies": [
        {
          "id": 1750058,
          "author_name": "beasterea1109",
          "author_url": "",
          "post_date": "2022-04-09T09:22:44.590000",
          "content": "<p>How come I saw this comment now,, <br>\nI came to a result that each of the species belongs to only one family, according to the train dataset.<br>\nIs this the same for the test dataset too??<br>\nThank You.</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 1778946,
          "author_name": "John Park",
          "author_url": "",
          "post_date": "2022-05-06T00:14:15.127000",
          "content": "<p>Hi <a href=\"https://www.kaggle.com/beasterea1109\" target=\"_blank\">@beasterea1109</a>!<br>\nYes, that is correct! Each species belongs to only one genus and one family.</p>",
          "votes": 0,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1735976": "taxonomy ranks are hierarchical, so each family have many genera, and each genus have many species.\n\neach given species will be a member of a single family and genus.",
    "1722597": "Yes, they are all included in the training data. Family and genus names are all unique, but species names are dependent on genus names. (For example, there could be same species name for different genus names, such as *Abies\tamabilis* vs *Cyperus amabilis*). ",
    "1722432": "I wonder if the \"category_id\" we have to predict for the test data is all included in the train data we are given. \nIf so, I was considering to check the dependency between the \"family\", \"genus\" and \"species\" to generate a boolean mask for sub classifiers for genus and species."
  }
}