{
  "id": 6232,
  "title": "Best solution with a 'regular' laptop",
  "url": "/competitions/expedia-personalized-sort/discussion/6232",
  "author_name": "",
  "post_date": "2013-11-05T11:03:12.640Z",
  "votes": null,
  "comment_count": 4,
  "views": 2772,
  "content": "<p>Well done to the winners. I am interested in the best leaderboard scores that can be achieved on a 'regular' laptop; say a new-ish laptop with 8GBs RAM, for example.</p>\n<p>I'm sure there must be some innovative approaches.</p>\n<p>Thanks.</p>",
  "messages": [
    {
      "id": "33253",
      "postDate": "11/05/2013 11:03:12",
      "content": "<p>Well done to the winners. I am interested in the best leaderboard scores that can be achieved on a 'regular' laptop; say a new-ish laptop with 8GBs RAM, for example.</p>\n<p>I'm sure there must be some innovative approaches.</p>\n<p>Thanks.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "33254",
      "postDate": "11/05/2013 11:09:10",
      "content": "<p>Until&nbsp;0.51381, I worked on 32 bits Windows (2 GB RAM). After that I switched to a laptop with 8 GB RAM and got&nbsp;0.52765 (after also improving features and techniques). I'm certain 0.52 is possible with just 2 GB RAM.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "33263",
      "postDate": "11/05/2013 13:36:51",
      "content": "<p>I was going to split the trainset in 10 parts, train 10 predictors and then average their result.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "33267",
      "postDate": "11/05/2013 14:05:29",
      "content": "<p>how you can do it ? can share your experience?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "33275",
      "postDate": "11/05/2013 15:49:30",
      "content": "<p>As an Expedia employee, I wasn't allowed to submit entries, however I did all my work on a Macbook Pro w/8GB ram. &nbsp;I submitted my results this morning (after the contest ended) for a score of .47497, which I thought was not bad for a first attempt at data science.</p>\n<p>I used an approach where I analyzed a single set of search results (as grouped by srch_id) at a time, and aggregated the results as I processed them, so I never had more than ~30 lines of the file in memory at any time. &nbsp;I was able to process the entire training file in ~40 seconds, and the test file in ~60 seconds.</p>\n<p>While the above score puts me squarely in the middle of the leader board, I attribute that mostly to my lack of knowledge on data analysis.</p>\n<p>This approach would work on a significantly lower-powered machine than the one I used, and I think a higher score could easily be obtained by somebody with more math knowledge :)</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 33254,
      "author_name": "gertjac",
      "author_url": "",
      "post_date": "11/05/2013 11:09:10",
      "content": "<p>Until&nbsp;0.51381, I worked on 32 bits Windows (2 GB RAM). After that I switched to a laptop with 8 GB RAM and got&nbsp;0.52765 (after also improving features and techniques). I'm certain 0.52 is possible with just 2 GB RAM.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 33263,
      "author_name": "ma2rten",
      "author_url": "",
      "post_date": "11/05/2013 13:36:51",
      "content": "<p>I was going to split the trainset in 10 parts, train 10 predictors and then average their result.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 33267,
      "author_name": "dylanfan1",
      "author_url": "",
      "post_date": "11/05/2013 14:05:29",
      "content": "<p>how you can do it ? can share your experience?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 33275,
      "author_name": "beaujackson",
      "author_url": "",
      "post_date": "11/05/2013 15:49:30",
      "content": "<p>As an Expedia employee, I wasn't allowed to submit entries, however I did all my work on a Macbook Pro w/8GB ram. &nbsp;I submitted my results this morning (after the contest ended) for a score of .47497, which I thought was not bad for a first attempt at data science.</p>\n<p>I used an approach where I analyzed a single set of search results (as grouped by srch_id) at a time, and aggregated the results as I processed them, so I never had more than ~30 lines of the file in memory at any time. &nbsp;I was able to process the entire training file in ~40 seconds, and the test file in ~60 seconds.</p>\n<p>While the above score puts me squarely in the middle of the leader board, I attribute that mostly to my lack of knowledge on data analysis.</p>\n<p>This approach would work on a significantly lower-powered machine than the one I used, and I think a higher score could easily be obtained by somebody with more math knowledge :)</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "33253": "",
    "33254": "",
    "33263": "",
    "33267": "",
    "33275": ""
  },
  "source": "meta"
}