{
  "id": 311134,
  "title": "Collaborative filter",
  "url": "/competitions/h-and-m-personalized-fashion-recommendations/discussion/311134",
  "author_name": "",
  "post_date": "2022-03-05T01:18:25.946144200Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Is collaborative filter feasible for this competition?<br>\nA few solutions are rule based so far.</p>\n<p>There are so many customers and articles. Not sure if we can score all of them using some algorithms on kaggle platform.</p>",
  "messages": [
    {
      "id": "1712484",
      "postDate": "03/05/2022 01:18:25",
      "content": "<p>Is collaborative filter feasible for this competition?<br>\nA few solutions are rule based so far.</p>\n<p>There are so many customers and articles. Not sure if we can score all of them using some algorithms on kaggle platform.</p>",
      "rawMarkdown": "Is collaborative filter feasible for this competition?\nA few solutions are rule based so far.\n\nThere are so many customers and articles. Not sure if we can score all of them using some algorithms on kaggle platform.",
      "votes": null
    },
    {
      "id": "1712764",
      "postDate": "03/05/2022 09:57:56",
      "content": "<p>For now, CF algorithms like ALS and others  are giving lower results than simple heuristics, I think that this is due to the fact that classical CF algorithms have a problem with sequential data (purchase history). Besides, we have a lot of info about customer and products that we are not taking into account, so that might help in future solutions.</p>",
      "rawMarkdown": "For now, CF algorithms like ALS and others  are giving lower results than simple heuristics, I think that this is due to the fact that classical CF algorithms have a problem with sequential data (purchase history). Besides, we have a lot of info about customer and products that we are not taking into account, so that might help in future solutions.",
      "votes": null
    },
    {
      "id": "1712942",
      "postDate": "03/05/2022 14:03:59",
      "content": "<p>You may be able to reduce memory by following this thread <a href=\"https://www.kaggle.com/c/h-and-m-personalized-fashion-recommendations/discussion/308635\" target=\"_blank\">https://www.kaggle.com/c/h-and-m-personalized-fashion-recommendations/discussion/308635</a></p>",
      "rawMarkdown": "You may be able to reduce memory by following this thread https://www.kaggle.com/c/h-and-m-personalized-fashion-recommendations/discussion/308635",
      "votes": null
    },
    {
      "id": "1714871",
      "postDate": "03/07/2022 12:40:06",
      "content": "<p>CF algorithms like Jaccard similarity it gets lose its accuracy while training huge customers data, I found the CF approach is not suggestable approach to perform for this project. This is my opinion</p>",
      "rawMarkdown": "CF algorithms like Jaccard similarity it gets lose its accuracy while training huge customers data, I found the CF approach is not suggestable approach to perform for this project. This is my opinion",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1712764,
      "author_name": "souamesannis",
      "author_url": "",
      "post_date": "03/05/2022 09:57:56",
      "content": "<p>For now, CF algorithms like ALS and others  are giving lower results than simple heuristics, I think that this is due to the fact that classical CF algorithms have a problem with sequential data (purchase history). Besides, we have a lot of info about customer and products that we are not taking into account, so that might help in future solutions.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1712942,
      "author_name": "atulverma",
      "author_url": "",
      "post_date": "03/05/2022 14:03:59",
      "content": "<p>You may be able to reduce memory by following this thread <a href=\"https://www.kaggle.com/c/h-and-m-personalized-fashion-recommendations/discussion/308635\" target=\"_blank\">https://www.kaggle.com/c/h-and-m-personalized-fashion-recommendations/discussion/308635</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1714871,
      "author_name": "arunkumar1809",
      "author_url": "",
      "post_date": "03/07/2022 12:40:06",
      "content": "<p>CF algorithms like Jaccard similarity it gets lose its accuracy while training huge customers data, I found the CF approach is not suggestable approach to perform for this project. This is my opinion</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1712484": "Is collaborative filter feasible for this competition?\nA few solutions are rule based so far.\n\nThere are so many customers and articles. Not sure if we can score all of them using some algorithms on kaggle platform.",
    "1712764": "For now, CF algorithms like ALS and others  are giving lower results than simple heuristics, I think that this is due to the fact that classical CF algorithms have a problem with sequential data (purchase history). Besides, we have a lot of info about customer and products that we are not taking into account, so that might help in future solutions.",
    "1712942": "You may be able to reduce memory by following this thread https://www.kaggle.com/c/h-and-m-personalized-fashion-recommendations/discussion/308635",
    "1714871": "CF algorithms like Jaccard similarity it gets lose its accuracy while training huge customers data, I found the CF approach is not suggestable approach to perform for this project. This is my opinion"
  },
  "source": "meta"
}