{
  "id": 52005,
  "title": "Introduce xLearn package for this competion",
  "url": "/competitions/talkingdata-adtracking-fraud-detection/discussion/52005",
  "author_name": "",
  "post_date": "2018-03-15T04:23:21.963209800Z",
  "votes": -20,
  "comment_count": 7,
  "views": 0,
  "content": "<p><a href=\"https://github.com/aksnzhy/xlearn\">https://github.com/aksnzhy/xlearn</a></p>\n\n<p>xLearn is a high performance, easy-to-use, and scalable machine learning package, which can be used to solve large-scale machine learning problems. xLearn is especially useful for solving machine learning problems on large-scale sparse data, which is very common in Internet services such as online advertisement and recommender systems in recent years. If you are the user of liblinear, libfm, or libffm, now the xLearn is your another better choice.</p>\n\n<p>We get the state-of-the-art score in previous CTR prediction competitions by using xLearn. </p>",
  "messages": [
    {
      "id": "296345",
      "postDate": "03/15/2018 04:23:21",
      "content": "<p><a href=\"https://github.com/aksnzhy/xlearn\">https://github.com/aksnzhy/xlearn</a></p>\n\n<p>xLearn is a high performance, easy-to-use, and scalable machine learning package, which can be used to solve large-scale machine learning problems. xLearn is especially useful for solving machine learning problems on large-scale sparse data, which is very common in Internet services such as online advertisement and recommender systems in recent years. If you are the user of liblinear, libfm, or libffm, now the xLearn is your another better choice.</p>\n\n<p>We get the state-of-the-art score in previous CTR prediction competitions by using xLearn. </p>",
      "rawMarkdown": "https://github.com/aksnzhy/xlearn\n\nxLearn is a high performance, easy-to-use, and scalable machine learning package, which can be used to solve large-scale machine learning problems. xLearn is especially useful for solving machine learning problems on large-scale sparse data, which is very common in Internet services such as online advertisement and recommender systems in recent years. If you are the user of liblinear, libfm, or libffm, now the xLearn is your another better choice.\n\nWe get the state-of-the-art score in previous CTR prediction competitions by using xLearn.",
      "votes": null
    },
    {
      "id": "296359",
      "postDate": "03/15/2018 04:54:19",
      "content": "<p>WOW, you have 15 same posts.....</p>",
      "rawMarkdown": "WOW, you have 15 same posts.....",
      "votes": null
    },
    {
      "id": "296380",
      "postDate": "03/15/2018 05:46:39",
      "content": "<p>Why not provide a starter code with xlearn in the kernels?</p>",
      "rawMarkdown": "Why not provide a starter code with xlearn in the kernels?",
      "votes": null
    },
    {
      "id": "296431",
      "postDate": "03/15/2018 07:47:03",
      "content": "<p>I would guess because it takes more time than just writing in a conversion rate competition how good it was \"in previous CTR  prediction competitions\".</p>",
      "rawMarkdown": "I would guess because it takes more time than just writing in a conversion rate competition how good it was \"in previous CTR  prediction competitions\".",
      "votes": null
    },
    {
      "id": "296539",
      "postDate": "03/15/2018 11:56:48",
      "content": "<p>Very good point @Asparuh.</p>",
      "rawMarkdown": "Very good point @Asparuh.",
      "votes": null
    },
    {
      "id": "296580",
      "postDate": "03/15/2018 13:17:58",
      "content": "<p>Write a kernel, run it and submit the result. If it gives a good score, everyone will use it.</p>",
      "rawMarkdown": "Write a kernel, run it and submit the result. If it gives a good score, everyone will use it.",
      "votes": null
    },
    {
      "id": "296589",
      "postDate": "03/15/2018 13:47:36",
      "content": "<p><img src=\"https://www.jiuwa.net/tuku/20170926/cNfX3smm.jpg\" alt=\"enter image description here\"></p>",
      "rawMarkdown": "![enter image description here][1]\n\n\n  [1]: https://www.jiuwa.net/tuku/20170926/cNfX3smm.jpg",
      "votes": null
    },
    {
      "id": "296621",
      "postDate": "03/15/2018 14:31:24",
      "content": "<p>dodolong - Please stop repeating these posts every contest. If it is posted again, I'll remove all of them.</p>",
      "rawMarkdown": "dodolong - Please stop repeating these posts every contest. If it is posted again, I'll remove all of them.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 296359,
      "author_name": "shujian",
      "author_url": "",
      "post_date": "03/15/2018 04:54:19",
      "content": "<p>WOW, you have 15 same posts.....</p>",
      "votes": null,
      "replies": [
        {
          "id": 296380,
          "author_name": "shujian",
          "author_url": "",
          "post_date": "03/15/2018 05:46:39",
          "content": "<p>Why not provide a starter code with xlearn in the kernels?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 296431,
          "author_name": "asparuhhristov",
          "author_url": "",
          "post_date": "03/15/2018 07:47:03",
          "content": "<p>I would guess because it takes more time than just writing in a conversion rate competition how good it was \"in previous CTR  prediction competitions\".</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 296539,
          "author_name": "shujian",
          "author_url": "",
          "post_date": "03/15/2018 11:56:48",
          "content": "<p>Very good point @Asparuh.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 296580,
      "author_name": "cczaixian",
      "author_url": "",
      "post_date": "03/15/2018 13:17:58",
      "content": "<p>Write a kernel, run it and submit the result. If it gives a good score, everyone will use it.</p>",
      "votes": null,
      "replies": [
        {
          "id": 296589,
          "author_name": "shujian",
          "author_url": "",
          "post_date": "03/15/2018 13:47:36",
          "content": "<p><img src=\"https://www.jiuwa.net/tuku/20170926/cNfX3smm.jpg\" alt=\"enter image description here\"></p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 296621,
      "author_name": "inversion",
      "author_url": "",
      "post_date": "03/15/2018 14:31:24",
      "content": "<p>dodolong - Please stop repeating these posts every contest. If it is posted again, I'll remove all of them.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "296345": "https://github.com/aksnzhy/xlearn\n\nxLearn is a high performance, easy-to-use, and scalable machine learning package, which can be used to solve large-scale machine learning problems. xLearn is especially useful for solving machine learning problems on large-scale sparse data, which is very common in Internet services such as online advertisement and recommender systems in recent years. If you are the user of liblinear, libfm, or libffm, now the xLearn is your another better choice.\n\nWe get the state-of-the-art score in previous CTR prediction competitions by using xLearn.",
    "296359": "WOW, you have 15 same posts.....",
    "296380": "Why not provide a starter code with xlearn in the kernels?",
    "296431": "I would guess because it takes more time than just writing in a conversion rate competition how good it was \"in previous CTR  prediction competitions\".",
    "296539": "Very good point @Asparuh.",
    "296580": "Write a kernel, run it and submit the result. If it gives a good score, everyone will use it.",
    "296589": "![enter image description here][1]\n\n\n  [1]: https://www.jiuwa.net/tuku/20170926/cNfX3smm.jpg",
    "296621": "dodolong - Please stop repeating these posts every contest. If it is posted again, I'll remove all of them."
  },
  "source": "meta"
}