{
  "id": 21981,
  "title": "Just curious about the running time",
  "url": "/competitions/facebook-v-predicting-check-ins/discussion/21981",
  "author_name": "",
  "post_date": "2016-06-30T19:02:44.243Z",
  "votes": null,
  "comment_count": 1,
  "views": 200,
  "content": "<p>I'm just curious about all your guys' running time.  </p>\n\n<p>This is my first time ever solving problems on Kaggle.  The naive knn results in score around 0.3 which is disappointing but it is really fast.  I was trying pruning+cluster+classifier per cluster but gave up since the running time was insane, by which I mean approx. 4-5 days.  Acutally my job was killed by server because of limited resources.</p>\n\n<p>So I just wonder what is your approx running time and memory usage.</p>",
  "messages": [
    {
      "id": "125598",
      "postDate": "06/30/2016 19:02:44",
      "content": "<p>I'm just curious about all your guys' running time.  </p>\n\n<p>This is my first time ever solving problems on Kaggle.  The naive knn results in score around 0.3 which is disappointing but it is really fast.  I was trying pruning+cluster+classifier per cluster but gave up since the running time was insane, by which I mean approx. 4-5 days.  Acutally my job was killed by server because of limited resources.</p>\n\n<p>So I just wonder what is your approx running time and memory usage.</p>",
      "rawMarkdown": "I'm just curious about all your guys' running time.  \r\n\r\nThis is my first time ever solving problems on Kaggle.  The naive knn results in score around 0.3 which is disappointing but it is really fast.  I was trying pruning+cluster+classifier per cluster but gave up since the running time was insane, by which I mean approx. 4-5 days.  Acutally my job was killed by server because of limited resources.\r\n\r\nSo I just wonder what is your approx running time and memory usage.",
      "votes": null
    },
    {
      "id": "125601",
      "postDate": "06/30/2016 19:25:17",
      "content": "<p>Welcome to Kaggle! :-)</p>\n\n<p>As you see in the script section the k-NN solutions can be calculated in less than 20 minutes with pretty decent results. The tree models (Random Forest, Extreme Gradient Boosting etc.) will take much longer, but may provide better accuracy. My longest run so far has been 48 hours. My patience doesn't last any longer. ;-)</p>\n\n<p>I assume everyone split the data into grids, so the memory consumption will be very low in this challenge. My latest submission took up 350 mb in R. </p>",
      "rawMarkdown": "Welcome to Kaggle! :-)\r\n\r\nAs you see in the script section the k-NN solutions can be calculated in less than 20 minutes with pretty decent results. The tree models (Random Forest, Extreme Gradient Boosting etc.) will take much longer, but may provide better accuracy. My longest run so far has been 48 hours. My patience doesn't last any longer. ;-)\r\n\r\nI assume everyone split the data into grids, so the memory consumption will be very low in this challenge. My latest submission took up 350 mb in R.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 125601,
      "author_name": "frederik",
      "author_url": "",
      "post_date": "06/30/2016 19:25:17",
      "content": "<p>Welcome to Kaggle! :-)</p>\n\n<p>As you see in the script section the k-NN solutions can be calculated in less than 20 minutes with pretty decent results. The tree models (Random Forest, Extreme Gradient Boosting etc.) will take much longer, but may provide better accuracy. My longest run so far has been 48 hours. My patience doesn't last any longer. ;-)</p>\n\n<p>I assume everyone split the data into grids, so the memory consumption will be very low in this challenge. My latest submission took up 350 mb in R. </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "125598": "I'm just curious about all your guys' running time.  \r\n\r\nThis is my first time ever solving problems on Kaggle.  The naive knn results in score around 0.3 which is disappointing but it is really fast.  I was trying pruning+cluster+classifier per cluster but gave up since the running time was insane, by which I mean approx. 4-5 days.  Acutally my job was killed by server because of limited resources.\r\n\r\nSo I just wonder what is your approx running time and memory usage.",
    "125601": "Welcome to Kaggle! :-)\r\n\r\nAs you see in the script section the k-NN solutions can be calculated in less than 20 minutes with pretty decent results. The tree models (Random Forest, Extreme Gradient Boosting etc.) will take much longer, but may provide better accuracy. My longest run so far has been 48 hours. My patience doesn't last any longer. ;-)\r\n\r\nI assume everyone split the data into grids, so the memory consumption will be very low in this challenge. My latest submission took up 350 mb in R."
  },
  "source": "meta"
}