{
  "id": 8133,
  "title": "Anyone using Mahout?",
  "url": "/competitions/acquire-valued-shoppers-challenge/discussion/8133",
  "author_name": "",
  "post_date": "2014-05-13T17:10:12.897Z",
  "votes": null,
  "comment_count": 4,
  "views": 2006,
  "content": "<p>Is anyone using Mahout for this challenge? &nbsp;If so what algorithm?</p>\n\n<p>I'm totally new to data science, so I'm learning. &nbsp;But I got a couple Mahout examples running, I don't think I'm applying the correct algorithm though to analyze the data.</p>",
  "messages": [
    {
      "id": "44473",
      "postDate": "05/13/2014 17:10:12",
      "content": "<p>Is anyone using Mahout for this challenge? &nbsp;If so what algorithm?</p>\n\n<p>I'm totally new to data science, so I'm learning. &nbsp;But I got a couple Mahout examples running, I don't think I'm applying the correct algorithm though to analyze the data.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "44480",
      "postDate": "05/13/2014 19:01:41",
      "content": "<p>Mahout&nbsp;can do&nbsp;logistic regression with SGD, which is quite similar to what&nbsp;most people here seem to use (https://mahout.apache.org/users/classification/logistic-regression.html).&nbsp;Any other classifier may work too though. Random forrest could&nbsp;be&nbsp;worth a try.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "48627",
      "postDate": "06/04/2014 13:26:01",
      "content": "<p>I am using mahout too. However for feature extraction, you will need to use a different set of tools - my preference would be something that runs on hadoop so that it can process the huge transaction data</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "48767",
      "postDate": "06/06/2014 11:01:42",
      "content": "<p>I'm using SVM that comes with <a href=\"http://spark.apache.org/\">spark</a>. Like mahout, it can run on top of Hadoop but it uses memory more efficiently.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "48779",
      "postDate": "06/06/2014 13:16:18",
      "content": "<p>Hey Paulo</p>\n\n<p>I try to learn Apache Spark for a university project. Maybe we could team up and I could help you to find some more features.&nbsp;</p>\n\n<p>You can write me&nbsp;under claudio.schroeder86@gmail.com&nbsp;. Would be great to share some ideas.</p>\n\n<p>Cheers,</p>\n<p>Claudio</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 44480,
      "author_name": "breuker",
      "author_url": "",
      "post_date": "05/13/2014 19:01:41",
      "content": "<p>Mahout&nbsp;can do&nbsp;logistic regression with SGD, which is quite similar to what&nbsp;most people here seem to use (https://mahout.apache.org/users/classification/logistic-regression.html).&nbsp;Any other classifier may work too though. Random forrest could&nbsp;be&nbsp;worth a try.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 48627,
      "author_name": "subbusw",
      "author_url": "",
      "post_date": "06/04/2014 13:26:01",
      "content": "<p>I am using mahout too. However for feature extraction, you will need to use a different set of tools - my preference would be something that runs on hadoop so that it can process the huge transaction data</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 48767,
      "author_name": "paulomagalhaes",
      "author_url": "",
      "post_date": "06/06/2014 11:01:42",
      "content": "<p>I'm using SVM that comes with <a href=\"http://spark.apache.org/\">spark</a>. Like mahout, it can run on top of Hadoop but it uses memory more efficiently.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 48779,
      "author_name": "claudios",
      "author_url": "",
      "post_date": "06/06/2014 13:16:18",
      "content": "<p>Hey Paulo</p>\n\n<p>I try to learn Apache Spark for a university project. Maybe we could team up and I could help you to find some more features.&nbsp;</p>\n\n<p>You can write me&nbsp;under claudio.schroeder86@gmail.com&nbsp;. Would be great to share some ideas.</p>\n\n<p>Cheers,</p>\n<p>Claudio</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "44473": "",
    "44480": "",
    "48627": "",
    "48767": "",
    "48779": ""
  },
  "source": "meta"
}