{
  "id": 9710,
  "title": "Ads containing phone/mail/url aren't blocked be default?",
  "url": "/competitions/avito-prohibited-content/discussion/9710",
  "author_name": "",
  "post_date": "2014-07-10T18:21:17.490Z",
  "votes": null,
  "comment_count": 1,
  "views": 880,
  "content": "<p>Hi!&nbsp;</p>\n<p>Avito rules say that&nbsp;ad cant contain phone/mail/url.</p>\n<p>http://www.avito.ru/help/osnovnye_prichiny_blokirovki_obyavlenij_i_uchyotnyh_zapisej</p>\n<p>But in the training data there are lots of not blocked ads with this info (f.e. id&nbsp;10000025).&nbsp;</p>\n<p>Is this right, that such ads must be blocked dy default by my algoritm, or i should ignore avito rules and base model only on training data?</p>",
  "messages": [
    {
      "id": "50305",
      "postDate": "07/10/2014 18:21:17",
      "content": "<p>Hi!&nbsp;</p>\n<p>Avito rules say that&nbsp;ad cant contain phone/mail/url.</p>\n<p>http://www.avito.ru/help/osnovnye_prichiny_blokirovki_obyavlenij_i_uchyotnyh_zapisej</p>\n<p>But in the training data there are lots of not blocked ads with this info (f.e. id&nbsp;10000025).&nbsp;</p>\n<p>Is this right, that such ads must be blocked dy default by my algoritm, or i should ignore avito rules and base model only on training data?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "50318",
      "postDate": "07/10/2014 21:49:01",
      "content": "<p>Hi Oleg,</p>\n<p>From a data science perspective, I suggest you build your model based on the training data. There may be other reasons for not blocking an ad even though it contains phone/email/url. Those ads may be special cases. It is a good idea to use that page&nbsp;in avito.ru as a guide, but your main source of information is the data itself.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 50318,
      "author_name": "luistandalla",
      "author_url": "",
      "post_date": "07/10/2014 21:49:01",
      "content": "<p>Hi Oleg,</p>\n<p>From a data science perspective, I suggest you build your model based on the training data. There may be other reasons for not blocking an ad even though it contains phone/email/url. Those ads may be special cases. It is a good idea to use that page&nbsp;in avito.ru as a guide, but your main source of information is the data itself.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "50305": "",
    "50318": ""
  },
  "source": "meta"
}