{
  "id": 13890,
  "title": "Python Binary Classification",
  "url": "/competitions/facebook-recruiting-iv-human-or-bot/discussion/13890",
  "author_name": "",
  "post_date": "2015-04-30T19:52:02.723Z",
  "votes": null,
  "comment_count": 7,
  "views": 3095,
  "content": "<p>I'm a college student interested in (finally) learning Python for this challenge. For almost all other competitions, I've used R exclusively.&nbsp;</p>\n\n<p>That said, what are the best resources to start learning&nbsp;Python?&nbsp;</p>\n<p>Additionally, what are good resources to start doing binary classification in Python?&nbsp;</p>\n\n<p>Thank you!</p>",
  "messages": [
    {
      "id": "75727",
      "postDate": "04/30/2015 19:52:02",
      "content": "<p>I'm a college student interested in (finally) learning Python for this challenge. For almost all other competitions, I've used R exclusively.&nbsp;</p>\n\n<p>That said, what are the best resources to start learning&nbsp;Python?&nbsp;</p>\n<p>Additionally, what are good resources to start doing binary classification in Python?&nbsp;</p>\n\n<p>Thank you!</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "75730",
      "postDate": "04/30/2015 19:56:03",
      "content": "<p>The two most commonly used packages are pandas and scikit-learn. You can make a lot of progress with only those two packages.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "75736",
      "postDate": "04/30/2015 20:05:10",
      "content": "<p>I am exactly in the opposite side :P, I use Python for most of my work and very few times R</p>\n<p>I would go with&nbsp;http://scikit-learn.org/stable/ you can <span style=\"text-decoration: line-through\">easily</span> get over 0.85 score (confirmed) with Scikit for sure.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "76567",
      "postDate": "05/04/2015 12:55:46",
      "content": "<p>@NxGTR i can able to train and test 1000 sample with scikitlearn, how to process whole sample population?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "76612",
      "postDate": "05/04/2015 15:30:27",
      "content": "<p>@preeth kumar I do not fully understand your question, are you running out-of-memory?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "76892",
      "postDate": "05/05/2015 03:54:33",
      "content": "<p>@NxGTR, thanks for prompt reply, i have only 4 GB of RAM in that how can i process 5 lakhs records of data. it takes so much time for even 1000 records .</p>\n\n<p>thats why i asked you how to process large data set in scikit learn python , whether we need to use any hadoop like framework.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "76920",
      "postDate": "05/05/2015 04:51:37",
      "content": "<p>This kind of competition (Jobs), does not allow much cooperation, and it makes sense, you will be interviewed individually, you need first to prove you are up for the challenge.</p>\n<p>I would suggest you to move to&nbsp;<a href=\"https://www.kaggle.com/c/data-science-london-scikit-learn\">this</a>&nbsp;competition, there you will find lots of good material, and lots of people would be able to help you, Facebook Recruiting IV&nbsp;might not be a good competition&nbsp;for entirely beginners.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "77362",
      "postDate": "05/06/2015 17:18:42",
      "content": "<p>@Greg, I learned Python for the last competition I competed in (TFI Restaurant) instead of using R (which I had tried to teach myself for the previous ones I submitted entries to).&nbsp; I find Python much easier to use, and I use Enthought Canopy (https://www.enthought.com/products/canopy/).&nbsp; The base version is free and includes a lot of libraries, plus it has a nice GUI for programming and running code (I'm not much of computer scientist and this made learning a lot easier for me).</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 75730,
      "author_name": "scku192636",
      "author_url": "",
      "post_date": "04/30/2015 19:56:03",
      "content": "<p>The two most commonly used packages are pandas and scikit-learn. You can make a lot of progress with only those two packages.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 75736,
      "author_name": "carloshuertas",
      "author_url": "",
      "post_date": "04/30/2015 20:05:10",
      "content": "<p>I am exactly in the opposite side :P, I use Python for most of my work and very few times R</p>\n<p>I would go with&nbsp;http://scikit-learn.org/stable/ you can <span style=\"text-decoration: line-through\">easily</span> get over 0.85 score (confirmed) with Scikit for sure.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 76567,
      "author_name": "preeth",
      "author_url": "",
      "post_date": "05/04/2015 12:55:46",
      "content": "<p>@NxGTR i can able to train and test 1000 sample with scikitlearn, how to process whole sample population?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 76612,
      "author_name": "carloshuertas",
      "author_url": "",
      "post_date": "05/04/2015 15:30:27",
      "content": "<p>@preeth kumar I do not fully understand your question, are you running out-of-memory?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 76892,
      "author_name": "preeth",
      "author_url": "",
      "post_date": "05/05/2015 03:54:33",
      "content": "<p>@NxGTR, thanks for prompt reply, i have only 4 GB of RAM in that how can i process 5 lakhs records of data. it takes so much time for even 1000 records .</p>\n\n<p>thats why i asked you how to process large data set in scikit learn python , whether we need to use any hadoop like framework.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 76920,
      "author_name": "carloshuertas",
      "author_url": "",
      "post_date": "05/05/2015 04:51:37",
      "content": "<p>This kind of competition (Jobs), does not allow much cooperation, and it makes sense, you will be interviewed individually, you need first to prove you are up for the challenge.</p>\n<p>I would suggest you to move to&nbsp;<a href=\"https://www.kaggle.com/c/data-science-london-scikit-learn\">this</a>&nbsp;competition, there you will find lots of good material, and lots of people would be able to help you, Facebook Recruiting IV&nbsp;might not be a good competition&nbsp;for entirely beginners.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 77362,
      "author_name": "thawatt",
      "author_url": "",
      "post_date": "05/06/2015 17:18:42",
      "content": "<p>@Greg, I learned Python for the last competition I competed in (TFI Restaurant) instead of using R (which I had tried to teach myself for the previous ones I submitted entries to).&nbsp; I find Python much easier to use, and I use Enthought Canopy (https://www.enthought.com/products/canopy/).&nbsp; The base version is free and includes a lot of libraries, plus it has a nice GUI for programming and running code (I'm not much of computer scientist and this made learning a lot easier for me).</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "75727": "",
    "75730": "",
    "75736": "",
    "76567": "",
    "76612": "",
    "76892": "",
    "76920": "",
    "77362": ""
  },
  "source": "meta"
}