{
  "id": 27170,
  "title": "Implicit Look-alike Modelling",
  "url": "/competitions/outbrain-click-prediction/discussion/27170",
  "author_name": "",
  "post_date": "2017-01-01T15:16:23.873Z",
  "votes": 2,
  "comment_count": 1,
  "views": 169,
  "content": "<p>Is anyone trying something like this?</p>\n\n<p><a href=\"https://arxiv.org/pdf/1601.02377v1.pdf\">Transfer Collaborative Filtering to CTR Estimation</a></p>\n\n<p>Summary:</p>\n\n<ul>\n<li>Web browsing prediction using FM: user x publisher* (page_views)</li>\n<li>CTR prediction using FM: weights of the user features and publisher* features in CTR task are assumed to be generated from the counterparts in CF task (as a prior)</li>\n</ul>\n\n<p>*publisher/document/whatever</p>",
  "messages": [
    {
      "id": "153445",
      "postDate": "01/01/2017 15:16:23",
      "content": "<p>Is anyone trying something like this?</p>\n\n<p><a href=\"https://arxiv.org/pdf/1601.02377v1.pdf\">Transfer Collaborative Filtering to CTR Estimation</a></p>\n\n<p>Summary:</p>\n\n<ul>\n<li>Web browsing prediction using FM: user x publisher* (page_views)</li>\n<li>CTR prediction using FM: weights of the user features and publisher* features in CTR task are assumed to be generated from the counterparts in CF task (as a prior)</li>\n</ul>\n\n<p>*publisher/document/whatever</p>",
      "rawMarkdown": "Is anyone trying something like this?\r\n\r\n[Transfer Collaborative Filtering to CTR Estimation][1]\r\n\r\nSummary:\r\n\r\n - Web browsing prediction using FM: user x publisher* (page_views)\r\n - CTR prediction using FM: weights of the user features and publisher* features in CTR task are assumed to be generated from the counterparts in CF task (as a prior)\r\n\r\n*publisher/document/whatever\r\n  [1]: https://arxiv.org/pdf/1601.02377v1.pdf",
      "votes": null
    },
    {
      "id": "1569624",
      "postDate": "11/03/2021 15:42:44",
      "content": "<p>Do we have any real world datasets which we can use to implement look-alike audiences/modelling?</p>",
      "rawMarkdown": "Do we have any real world datasets which we can use to implement look-alike audiences/modelling?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1569624,
      "author_name": "anandyati",
      "author_url": "",
      "post_date": "11/03/2021 15:42:44",
      "content": "<p>Do we have any real world datasets which we can use to implement look-alike audiences/modelling?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "153445": "Is anyone trying something like this?\r\n\r\n[Transfer Collaborative Filtering to CTR Estimation][1]\r\n\r\nSummary:\r\n\r\n - Web browsing prediction using FM: user x publisher* (page_views)\r\n - CTR prediction using FM: weights of the user features and publisher* features in CTR task are assumed to be generated from the counterparts in CF task (as a prior)\r\n\r\n*publisher/document/whatever\r\n  [1]: https://arxiv.org/pdf/1601.02377v1.pdf",
    "1569624": "Do we have any real world datasets which we can use to implement look-alike audiences/modelling?"
  },
  "source": "meta"
}