{
  "id": 294798,
  "title": "Bag of Words to combine description & tags",
  "url": "/competitions/cs5785-fall-2021-final/discussion/294798",
  "author_name": "",
  "post_date": "2021-12-12T22:08:50.341122300Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hello, I was wondering if anyone has experimented with different ways to combine description set and the tags to the images? For our approach we tried building a Bag of Words using either the descriptions or the tags (Thx <a href=\"https://www.kaggle.com/Clare\" target=\"_blank\">@Clare</a> for posting about it earlier) and use that to featurize the other. Unfortunately, the results of this approach aren't as accurate as we'd hoped. If someone can point out some of the mistakes in logic or approach that they ran into with this approach?</p>",
  "messages": [
    {
      "id": "1615899",
      "postDate": "12/12/2021 22:08:50",
      "content": "<p>Hello, I was wondering if anyone has experimented with different ways to combine description set and the tags to the images? For our approach we tried building a Bag of Words using either the descriptions or the tags (Thx <a href=\"https://www.kaggle.com/Clare\" target=\"_blank\">@Clare</a> for posting about it earlier) and use that to featurize the other. Unfortunately, the results of this approach aren't as accurate as we'd hoped. If someone can point out some of the mistakes in logic or approach that they ran into with this approach?</p>",
      "rawMarkdown": "Hello, I was wondering if anyone has experimented with different ways to combine description set and the tags to the images? For our approach we tried building a Bag of Words using either the descriptions or the tags (Thx @Clare for posting about it earlier) and use that to featurize the other. Unfortunately, the results of this approach aren't as accurate as we'd hoped. If someone can point out some of the mistakes in logic or approach that they ran into with this approach?",
      "votes": null
    },
    {
      "id": "1616010",
      "postDate": "12/13/2021 03:58:40",
      "content": "<p>you can't combine descriptions with tags. If you look into the testing set, the tags and descriptions don't match at all, the tags are with images.</p>",
      "rawMarkdown": "you can't combine descriptions with tags. If you look into the testing set, the tags and descriptions don't match at all, the tags are with images.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1616010,
      "author_name": "carolineming",
      "author_url": "",
      "post_date": "12/13/2021 03:58:40",
      "content": "<p>you can't combine descriptions with tags. If you look into the testing set, the tags and descriptions don't match at all, the tags are with images.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1615899": "Hello, I was wondering if anyone has experimented with different ways to combine description set and the tags to the images? For our approach we tried building a Bag of Words using either the descriptions or the tags (Thx @Clare for posting about it earlier) and use that to featurize the other. Unfortunately, the results of this approach aren't as accurate as we'd hoped. If someone can point out some of the mistakes in logic or approach that they ran into with this approach?",
    "1616010": "you can't combine descriptions with tags. If you look into the testing set, the tags and descriptions don't match at all, the tags are with images."
  },
  "source": "meta"
}