{
  "id": 51914,
  "title": "Let the hunt begin",
  "url": "/competitions/talkingdata-adtracking-fraud-detection/discussion/51914",
  "author_name": "",
  "post_date": "2018-03-14T09:28:35.864071900Z",
  "votes": 3,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I wonder given the distant past (a competition from 2 years ago), the near past (the initial test data) and the present (the IPs), how much time will people invest in finding a leakage in the data rather than building a \"reasonable\" model?\nI personally approach this competition as a great learning opportunity so I don't care so much if (whenever...) there is a leakage found. God help those striving for top positions in the LB...</p>",
  "messages": [
    {
      "id": "295839",
      "postDate": "03/14/2018 09:28:35",
      "content": "<p>I wonder given the distant past (a competition from 2 years ago), the near past (the initial test data) and the present (the IPs), how much time will people invest in finding a leakage in the data rather than building a \"reasonable\" model?\nI personally approach this competition as a great learning opportunity so I don't care so much if (whenever...) there is a leakage found. God help those striving for top positions in the LB...</p>",
      "rawMarkdown": "I wonder given the distant past (a competition from 2 years ago), the near past (the initial test data) and the present (the IPs), how much time will people invest in finding a leakage in the data rather than building a \"reasonable\" model?\nI personally approach this competition as a great learning opportunity so I don't care so much if (whenever...) there is a leakage found. God help those striving for top positions in the LB...",
      "votes": null
    },
    {
      "id": "297104",
      "postDate": "03/16/2018 09:27:24",
      "content": "<p>Hi ! For a beginner (like me) in Data Science it's a really great opportunity to learn ML because it's a classification problem and data structure is not too much elaborate !</p>\n\n<p>I choose to try to use some Machine Learning algorithms like Boosting it's really interesting to discover ML !  </p>",
      "rawMarkdown": "Hi ! For a beginner (like me) in Data Science it's a really great opportunity to learn ML because it's a classification problem and data structure is not too much elaborate !\n\nI choose to try to use some Machine Learning algorithms like Boosting it's really interesting to discover ML !",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 297104,
      "author_name": "nathanlauga",
      "author_url": "",
      "post_date": "03/16/2018 09:27:24",
      "content": "<p>Hi ! For a beginner (like me) in Data Science it's a really great opportunity to learn ML because it's a classification problem and data structure is not too much elaborate !</p>\n\n<p>I choose to try to use some Machine Learning algorithms like Boosting it's really interesting to discover ML !  </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "295839": "I wonder given the distant past (a competition from 2 years ago), the near past (the initial test data) and the present (the IPs), how much time will people invest in finding a leakage in the data rather than building a \"reasonable\" model?\nI personally approach this competition as a great learning opportunity so I don't care so much if (whenever...) there is a leakage found. God help those striving for top positions in the LB...",
    "297104": "Hi ! For a beginner (like me) in Data Science it's a really great opportunity to learn ML because it's a classification problem and data structure is not too much elaborate !\n\nI choose to try to use some Machine Learning algorithms like Boosting it's really interesting to discover ML !"
  },
  "source": "meta"
}