{
  "id": 5778,
  "title": "Memory issue with R",
  "url": "/competitions/expedia-personalized-sort/discussion/5778",
  "author_name": "Florent Renucci",
  "post_date": "2013-09-16T21:36:54.390000",
  "votes": 1,
  "comment_count": 0,
  "views": 884,
  "content": "<p>Anyone has already dealt with that issue ? &quot;Error : cannot allocate vector of size ... Gb&quot;</p>\n<p>&nbsp;</p>\n<p>The ff and ffbase package help importing large datasets, but as soon as we want to use &quot;regular&quot; packages/functions/algorithms such as Boosting, SVM or Random Forest, it is working only with small subsets of the training dataset, even with lots of RAM on a 64bits version of Windows and R. One post on Stack Overflow proposed to change the memory limit &quot;memory.limit(size = whatever)&quot;, but this is not enough.&nbsp;</p>\n<p>Any suggestion would be very appreciated ! :)</p>\n<p>Flo</p>",
  "messages": [
    {
      "id": 31012,
      "postDate": "2013-09-16T21:36:54.390Z",
      "content": "<p>Anyone has already dealt with that issue ? &quot;Error : cannot allocate vector of size ... Gb&quot;</p>\n<p>&nbsp;</p>\n<p>The ff and ffbase package help importing large datasets, but as soon as we want to use &quot;regular&quot; packages/functions/algorithms such as Boosting, SVM or Random Forest, it is working only with small subsets of the training dataset, even with lots of RAM on a 64bits version of Windows and R. One post on Stack Overflow proposed to change the memory limit &quot;memory.limit(size = whatever)&quot;, but this is not enough.&nbsp;</p>\n<p>Any suggestion would be very appreciated ! :)</p>\n<p>Flo</p>",
      "votes": 1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "31012": ""
  }
}