{
  "id": 23493,
  "title": "Train vs Test data",
  "url": "/competitions/bosch-production-line-performance/discussion/23493",
  "author_name": "",
  "post_date": "2016-09-08T13:43:37.583Z",
  "votes": null,
  "comment_count": 4,
  "views": 630,
  "content": "<p>Hi,\nI am new to this Kaggle I would like to understand what is the difference between Train and Test data?</p>\n\n<p>Thanks,\nTiru</p>",
  "messages": [
    {
      "id": "134794",
      "postDate": "09/08/2016 13:43:37",
      "content": "<p>Hi,\nI am new to this Kaggle I would like to understand what is the difference between Train and Test data?</p>\n\n<p>Thanks,\nTiru</p>",
      "rawMarkdown": "Hi,\r\nI am new to this Kaggle I would like to understand what is the difference between Train and Test data?\r\n\r\nThanks,\r\nTiru",
      "votes": null
    },
    {
      "id": "134795",
      "postDate": "09/08/2016 13:55:08",
      "content": "<p>Hi Tiru,</p>\n\n<p>Train data (train_numeric.csv) have response variable. Test data don't have it.<br>\nYou build model from train data and predict response variable for test data.</p>",
      "rawMarkdown": "Hi Tiru,\r\n\r\nTrain data (train_numeric.csv) have response variable. Test data don't have it.<br>\r\nYou build model from train data and predict response variable for test data.",
      "votes": null
    },
    {
      "id": "134817",
      "postDate": "09/08/2016 17:16:37",
      "content": "<p>Thank you Toshi</p>",
      "rawMarkdown": "Thank you Toshi",
      "votes": null
    },
    {
      "id": "134838",
      "postDate": "09/08/2016 19:26:11",
      "content": "<p>Tiru,</p>\n\n<p>My suggestion would be to check out the Titanic: Machine Learning from Disaster Competition.  It is a great place to get started and will walk you through all of the (many many many) intermediate steps that happen to train your model on the training data set and eventually use that model to predict the values on the test data set.</p>\n\n<p>This competition probably isn't a great place to get your feet wet given the size of the data, small proportion of positive responses, and ambiguity in what the features are.</p>\n\n<p>Good luck!</p>",
      "rawMarkdown": "Tiru,\r\n\r\nMy suggestion would be to check out the Titanic: Machine Learning from Disaster Competition.  It is a great place to get started and will walk you through all of the (many many many) intermediate steps that happen to train your model on the training data set and eventually use that model to predict the values on the test data set.\r\n\r\nThis competition probably isn't a great place to get your feet wet given the size of the data, small proportion of positive responses, and ambiguity in what the features are.\r\n\r\nGood luck!",
      "votes": null
    },
    {
      "id": "134844",
      "postDate": "09/08/2016 19:56:25",
      "content": "<p>Thank you Mathand will check it.</p>",
      "rawMarkdown": "Thank you Mathand will check it.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 134795,
      "author_name": "toshik",
      "author_url": "",
      "post_date": "09/08/2016 13:55:08",
      "content": "<p>Hi Tiru,</p>\n\n<p>Train data (train_numeric.csv) have response variable. Test data don't have it.<br>\nYou build model from train data and predict response variable for test data.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 134817,
      "author_name": "tirureddy",
      "author_url": "",
      "post_date": "09/08/2016 17:16:37",
      "content": "<p>Thank you Toshi</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 134838,
      "author_name": "mathandbikes",
      "author_url": "",
      "post_date": "09/08/2016 19:26:11",
      "content": "<p>Tiru,</p>\n\n<p>My suggestion would be to check out the Titanic: Machine Learning from Disaster Competition.  It is a great place to get started and will walk you through all of the (many many many) intermediate steps that happen to train your model on the training data set and eventually use that model to predict the values on the test data set.</p>\n\n<p>This competition probably isn't a great place to get your feet wet given the size of the data, small proportion of positive responses, and ambiguity in what the features are.</p>\n\n<p>Good luck!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 134844,
      "author_name": "tirureddy",
      "author_url": "",
      "post_date": "09/08/2016 19:56:25",
      "content": "<p>Thank you Mathand will check it.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "134794": "Hi,\r\nI am new to this Kaggle I would like to understand what is the difference between Train and Test data?\r\n\r\nThanks,\r\nTiru",
    "134795": "Hi Tiru,\r\n\r\nTrain data (train_numeric.csv) have response variable. Test data don't have it.<br>\r\nYou build model from train data and predict response variable for test data.",
    "134817": "Thank you Toshi",
    "134838": "Tiru,\r\n\r\nMy suggestion would be to check out the Titanic: Machine Learning from Disaster Competition.  It is a great place to get started and will walk you through all of the (many many many) intermediate steps that happen to train your model on the training data set and eventually use that model to predict the values on the test data set.\r\n\r\nThis competition probably isn't a great place to get your feet wet given the size of the data, small proportion of positive responses, and ambiguity in what the features are.\r\n\r\nGood luck!",
    "134844": "Thank you Mathand will check it."
  },
  "source": "meta"
}