{
  "id": 21200,
  "title": "Submission?",
  "url": "/competitions/facebook-v-predicting-check-ins/discussion/21200",
  "author_name": "",
  "post_date": "2016-05-25T06:01:46.727Z",
  "votes": null,
  "comment_count": 1,
  "views": 342,
  "content": "<p>Hi guys,</p>\n\n<p>Can anyone explain how we are getting 3 predictions for each row, I read some of the posts which talks about using the MAP concept and ml_metrics package, but I am still not able to understand how is it actually being implemented in the python code, suppose I am using KNN algorithm which predict one place_id for each row then how  can I make it predict three separate ids?</p>\n\n<p>Thanks</p>",
  "messages": [
    {
      "id": "121245",
      "postDate": "05/25/2016 06:01:46",
      "content": "<p>Hi guys,</p>\n\n<p>Can anyone explain how we are getting 3 predictions for each row, I read some of the posts which talks about using the MAP concept and ml_metrics package, but I am still not able to understand how is it actually being implemented in the python code, suppose I am using KNN algorithm which predict one place_id for each row then how  can I make it predict three separate ids?</p>\n\n<p>Thanks</p>",
      "rawMarkdown": "Hi guys,\r\n\r\nCan anyone explain how we are getting 3 predictions for each row, I read some of the posts which talks about using the MAP concept and ml_metrics package, but I am still not able to understand how is it actually being implemented in the python code, suppose I am using KNN algorithm which predict one place_id for each row then how  can I make it predict three separate ids?\r\n\r\nThanks",
      "votes": null
    },
    {
      "id": "121437",
      "postDate": "05/26/2016 09:33:26",
      "content": "<p>Besides getting a single predicted class, some KNN algorithms such as the one in sklearn also allows you to get the probability for each class. Then you may extract the top 3 classes with the highest probability.</p>",
      "rawMarkdown": "Besides getting a single predicted class, some KNN algorithms such as the one in sklearn also allows you to get the probability for each class. Then you may extract the top 3 classes with the highest probability.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 121437,
      "author_name": "citynight",
      "author_url": "",
      "post_date": "05/26/2016 09:33:26",
      "content": "<p>Besides getting a single predicted class, some KNN algorithms such as the one in sklearn also allows you to get the probability for each class. Then you may extract the top 3 classes with the highest probability.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "121245": "Hi guys,\r\n\r\nCan anyone explain how we are getting 3 predictions for each row, I read some of the posts which talks about using the MAP concept and ml_metrics package, but I am still not able to understand how is it actually being implemented in the python code, suppose I am using KNN algorithm which predict one place_id for each row then how  can I make it predict three separate ids?\r\n\r\nThanks",
    "121437": "Besides getting a single predicted class, some KNN algorithms such as the one in sklearn also allows you to get the probability for each class. Then you may extract the top 3 classes with the highest probability."
  },
  "source": "meta"
}