{
  "id": 10789,
  "title": "Confusions from Q&A",
  "url": "/competitions/avazu-ctr-prediction/discussion/10789",
  "author_name": "Need for jobs",
  "post_date": "2014-10-29T17:11:40.757000",
  "votes": 2,
  "comment_count": 0,
  "views": 752,
  "content": "<p>&quot;Perhaps there is no strong correlation between sampled CTR and true CTR with respect to time features.&quot;</p>\n<p>Hmm, we are supposed to predict the sampled CTR. If there's no strong correlation between them, e.g. by using a different subsampling ratio every day/hour, what can we say about the sampled CTR on the prediction day?</p>\n<p style=\"padding-left: 30px\"><em>2. Q: By the way, are there any explanations why CTR is maximal on Wednesday?</em></p>\n<p style=\"padding-left: 30px\"><em>A: We made the data containing 200k records per hour, that is, we first sampled some clicked records, and then added non-clicked records to make them adding up to 200k, so perhaps there is no strong correlation between sampled CTR and true CTR with respect to time features</em></p>",
  "messages": [
    {
      "id": 57002,
      "postDate": "2014-10-29T17:11:40.757Z",
      "content": "<p>&quot;Perhaps there is no strong correlation between sampled CTR and true CTR with respect to time features.&quot;</p>\n<p>Hmm, we are supposed to predict the sampled CTR. If there's no strong correlation between them, e.g. by using a different subsampling ratio every day/hour, what can we say about the sampled CTR on the prediction day?</p>\n<p style=\"padding-left: 30px\"><em>2. Q: By the way, are there any explanations why CTR is maximal on Wednesday?</em></p>\n<p style=\"padding-left: 30px\"><em>A: We made the data containing 200k records per hour, that is, we first sampled some clicked records, and then added non-clicked records to make them adding up to 200k, so perhaps there is no strong correlation between sampled CTR and true CTR with respect to time features</em></p>",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "57002": ""
  }
}