{
  "id": 313967,
  "title": "How to use the images?",
  "url": "/competitions/h-and-m-personalized-fashion-recommendations/discussion/313967",
  "author_name": "",
  "post_date": "2022-03-20T03:29:25.174746900Z",
  "votes": 6,
  "comment_count": 4,
  "views": 0,
  "content": "<p>We have lots of images in this competition (29GB), but how to make the most of the image information.</p>",
  "messages": [
    {
      "id": "1729406",
      "postDate": "03/20/2022 03:29:25",
      "content": "<p>We have lots of images in this competition (29GB), but how to make the most of the image information.</p>",
      "rawMarkdown": "We have lots of images in this competition (29GB), but how to make the most of the image information.",
      "votes": null
    },
    {
      "id": "1730124",
      "postDate": "03/21/2022 00:22:35",
      "content": "<p>There were many discussions as well as notebooks shared on this topic.</p>",
      "rawMarkdown": "There were many discussions as well as notebooks shared on this topic.",
      "votes": null
    },
    {
      "id": "1731133",
      "postDate": "03/22/2022 02:33:35",
      "content": "<p>The basic idea is to find similar images from CNN embeddings. Then there are at least two ways to use this. (1) Find similar customers who purchase similar items then recommend to one customer what the other customer is buying (2) Based on a customer's past purchases find similar images. Then recommend to that customer the items which are similar.</p>",
      "rawMarkdown": "The basic idea is to find similar images from CNN embeddings. Then there are at least two ways to use this. (1) Find similar customers who purchase similar items then recommend to one customer what the other customer is buying (2) Based on a customer's past purchases find similar images. Then recommend to that customer the items which are similar.",
      "votes": null
    },
    {
      "id": "1731603",
      "postDate": "03/22/2022 14:31:54",
      "content": "<p>There are some notebooks on this, but it is a very interesting question. It could certainly be done without using the images at all, so we will see how vital the images are to obtain an accurate prediction of future purchases.</p>",
      "rawMarkdown": "There are some notebooks on this, but it is a very interesting question. It could certainly be done without using the images at all, so we will see how vital the images are to obtain an accurate prediction of future purchases.",
      "votes": null
    },
    {
      "id": "1731724",
      "postDate": "03/22/2022 16:17:38",
      "content": "<p>Maybe it is not a question of either one or the other but use all available data and options depending upon your time and other resources </p>",
      "rawMarkdown": "Maybe it is not a question of either one or the other but use all available data and options depending upon your time and other resources",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1730124,
      "author_name": "atulverma",
      "author_url": "",
      "post_date": "03/21/2022 00:22:35",
      "content": "<p>There were many discussions as well as notebooks shared on this topic.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1731133,
      "author_name": "cdeotte",
      "author_url": "",
      "post_date": "03/22/2022 02:33:35",
      "content": "<p>The basic idea is to find similar images from CNN embeddings. Then there are at least two ways to use this. (1) Find similar customers who purchase similar items then recommend to one customer what the other customer is buying (2) Based on a customer's past purchases find similar images. Then recommend to that customer the items which are similar.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1731603,
      "author_name": "peterwarren",
      "author_url": "",
      "post_date": "03/22/2022 14:31:54",
      "content": "<p>There are some notebooks on this, but it is a very interesting question. It could certainly be done without using the images at all, so we will see how vital the images are to obtain an accurate prediction of future purchases.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1731724,
          "author_name": "atulverma",
          "author_url": "",
          "post_date": "03/22/2022 16:17:38",
          "content": "<p>Maybe it is not a question of either one or the other but use all available data and options depending upon your time and other resources </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1729406": "We have lots of images in this competition (29GB), but how to make the most of the image information.",
    "1730124": "There were many discussions as well as notebooks shared on this topic.",
    "1731133": "The basic idea is to find similar images from CNN embeddings. Then there are at least two ways to use this. (1) Find similar customers who purchase similar items then recommend to one customer what the other customer is buying (2) Based on a customer's past purchases find similar images. Then recommend to that customer the items which are similar.",
    "1731603": "There are some notebooks on this, but it is a very interesting question. It could certainly be done without using the images at all, so we will see how vital the images are to obtain an accurate prediction of future purchases.",
    "1731724": "Maybe it is not a question of either one or the other but use all available data and options depending upon your time and other resources"
  },
  "source": "meta"
}