![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2976852%2F7d82b142e44a9eed371e82ccbf81083e%2F2020-03-26%2014_51_41-Data%20Exploration%20_%20Kaggle.png?generation=1585231055016092&alt=media\")
On the \"Data\" section, it's only mentioned that this file contains fine-grained attributes description but not explained the attributes of the given JSON. Here is one category from the JSON file:
\n\n\n\n\n\n
{'id': 335, 'name': 'cheetah', 'supercategory': 'animal', 'level': 2}
Here id and name are self explanatory but what is supercategory and level?
On the \"Data\" section, it's only mentioned that this file contains fine-grained attributes description but not explained the attributes of the given JSON. Here is one category from the JSON file:
\n\n\n\n\n\n
{'id': 335, 'name': 'cheetah', 'supercategory': 'animal', 'level': 2}
Here id and name are self explanatory but what is supercategory and level?
Hey,
\n\nas you can see in this picture the supercategory describes every attribute a little further, for example, id 73(flamenco) and 25(mao) have the supercategory \"nickname\".
\n\n
So every attribute is grouped by a supercategory.\nRegarding the level feature, I guess, and this is only a guess that it has something to do with the following picture.
\n\n%202020%20at%20FGVC7%20_%20Kaggle.png?generation=1585231146992862&alt=media)
Maybe the level describes the depth of an attribute in these trees. I haven't checked that yet.
\n\nI hope I was able to help. If I'm mistaken somewhere please correct me.
\n\nI wish everyone a nice day :)
", "rawMarkdown": "Hey,\n\nas you can see in this picture the supercategory describes every attribute a little further, for example, id 73(flamenco) and 25(mao) have the supercategory \"nickname\".\n\n\n\nSo every attribute is grouped by a supercategory.\nRegarding the level feature, I guess, and this is only a guess that it has something to do with the following picture.\n\n%202020%20at%20FGVC7%20_%20Kaggle.png?generation=1585231146992862&alt=media)\n\nMaybe the level describes the depth of an attribute in these trees. I haven't checked that yet.\n\nI hope I was able to help. If I'm mistaken somewhere please correct me.\n\nI wish everyone a nice day :)", "votes": null }, { "id": "787989", "postDate": "03/27/2020 09:30:41", "content": "@jone93 So where is the description of classId? In this json files the \"id\" field seems to be the attributes.
", "rawMarkdown": "jone93 So where is the description of classId? In this json files the \"id\" field seems to be the attributes.", "votes": null }, { "id": "788131", "postDate": "03/27/2020 12:01:21", "content": "The classId in the train.csv corresponds to the id of the categories dict in the label_description.json
", "rawMarkdown": "The classId in the train.csv corresponds to the id of the categories dict in the label_description.json", "votes": null }, { "id": "788287", "postDate": "03/27/2020 14:43:53", "content": "@jone93 then what about attribute id?
", "rawMarkdown": "jone93 then what about attribute id?", "votes": null }, { "id": "788427", "postDate": "03/27/2020 16:55:21", "content": "The description of AttributeID and ClassID are both in the description JSON file. Check my Notebook, maybe it answers your question :)
\n\nhttps://www.kaggle.com/jone93/data-exploration
", "rawMarkdown": "The description of AttributeID and ClassID are both in the description JSON file. Check my Notebook, maybe it answers your question :)\n\nhttps://www.kaggle.com/jone93/data-exploration", "votes": null }, { "id": "800345", "postDate": "04/07/2020 10:17:17", "content": "Hey @jone93, does this super category information help the model in anyway? Does the model consider what a class's super category is while training or testing?
", "rawMarkdown": "Hey @jone93, does this super category information help the model in anyway? Does the model consider what a class's super category is while training or testing?", "votes": null }, { "id": "800416", "postDate": "04/07/2020 11:47:03", "content": "its really good!
", "rawMarkdown": "its really good!", "votes": null } ], "comments": [ { "id": 787072, "author_name": "jone93", "author_url": "", "post_date": "03/26/2020 14:01:53", "content": "Hey,
\n\nas you can see in this picture the supercategory describes every attribute a little further, for example, id 73(flamenco) and 25(mao) have the supercategory \"nickname\".
\n\n
So every attribute is grouped by a supercategory.\nRegarding the level feature, I guess, and this is only a guess that it has something to do with the following picture.
\n\n%202020%20at%20FGVC7%20_%20Kaggle.png?generation=1585231146992862&alt=media)
Maybe the level describes the depth of an attribute in these trees. I haven't checked that yet.
\n\nI hope I was able to help. If I'm mistaken somewhere please correct me.
\n\nI wish everyone a nice day :)
", "votes": null, "replies": [ { "id": 787989, "author_name": "kaushal2896", "author_url": "", "post_date": "03/27/2020 09:30:41", "content": "@jone93 So where is the description of classId? In this json files the \"id\" field seems to be the attributes.
", "votes": null, "replies": [] }, { "id": 788131, "author_name": "jone93", "author_url": "", "post_date": "03/27/2020 12:01:21", "content": "The classId in the train.csv corresponds to the id of the categories dict in the label_description.json
", "votes": null, "replies": [] }, { "id": 788287, "author_name": "kaushal2896", "author_url": "", "post_date": "03/27/2020 14:43:53", "content": "@jone93 then what about attribute id?
", "votes": null, "replies": [] }, { "id": 788427, "author_name": "jone93", "author_url": "", "post_date": "03/27/2020 16:55:21", "content": "The description of AttributeID and ClassID are both in the description JSON file. Check my Notebook, maybe it answers your question :)
\n\nhttps://www.kaggle.com/jone93/data-exploration
", "votes": null, "replies": [] }, { "id": 800345, "author_name": "rohitmidha23", "author_url": "", "post_date": "04/07/2020 10:17:17", "content": "Hey @jone93, does this super category information help the model in anyway? Does the model consider what a class's super category is while training or testing?
", "votes": null, "replies": [] } ] }, { "id": 800416, "author_name": "ganapathisubramanyam", "author_url": "", "post_date": "04/07/2020 11:47:03", "content": "its really good!
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "786726": "On the \"Data\" section, it's only mentioned that this file contains fine-grained attributes description but not explained the attributes of the given JSON. Here is one category from the JSON file: \n\n>`{'id': 335, 'name': 'cheetah', 'supercategory': 'animal', 'level': 2}`\n\nHere `id` and `name` are self explanatory but what is `supercategory` and `level`?", "787072": "Hey,\n\nas you can see in this picture the supercategory describes every attribute a little further, for example, id 73(flamenco) and 25(mao) have the supercategory \"nickname\".\n\n\n\nSo every attribute is grouped by a supercategory.\nRegarding the level feature, I guess, and this is only a guess that it has something to do with the following picture.\n\n%202020%20at%20FGVC7%20_%20Kaggle.png?generation=1585231146992862&alt=media)\n\nMaybe the level describes the depth of an attribute in these trees. I haven't checked that yet.\n\nI hope I was able to help. If I'm mistaken somewhere please correct me.\n\nI wish everyone a nice day :)", "787989": "jone93 So where is the description of classId? In this json files the \"id\" field seems to be the attributes.", "788131": "The classId in the train.csv corresponds to the id of the categories dict in the label_description.json", "788287": "jone93 then what about attribute id?", "788427": "The description of AttributeID and ClassID are both in the description JSON file. Check my Notebook, maybe it answers your question :)\n\nhttps://www.kaggle.com/jone93/data-exploration", "800345": "Hey @jone93, does this super category information help the model in anyway? Does the model consider what a class's super category is while training or testing?", "800416": "its really good!" }, "source": "meta" }