{
  "id": 4636,
  "title": "Actual values of test set",
  "url": "/competitions/digit-recognizer/discussion/4636",
  "author_name": "miya_clark",
  "post_date": "2013-05-20T15:11:41.323000",
  "votes": 0,
  "comment_count": 3,
  "views": null,
  "content": "<p>Hi,</p>\r\n<p>Does anybody know where can I find the actual results of the Test set?</p>\r\n<p>In the Data section it is stated that test set is actually the training set, but when I tried to match training and test set by rows they contain, I don't get any match.</p>\r\n<p>And second the size of training set is different than test set.</p>\r\n<p></p>\r\n<p>I would appreciate if anyone can tell me how can I test my own prediction results.</p>\r\n<p></p>\r\n<p></p>\r\n<p>Best,</p>\r\n<p>Miya</p>\r\n<p></p>",
  "messages": [
    {
      "id": 24536,
      "postDate": "2013-05-20T19:49:29.027Z",
      "content": "<p>You can split given training data and for instance use 80% of it for training, and 20% for tests.&nbsp;</p>"
    },
    {
      "id": 24532,
      "postDate": "2013-05-20T19:05:21.087Z",
      "content": "<p>Hi,</p>\r\n<p></p>\r\n<p>Thank you for you answer. But If I want to estimate misclassification&nbsp;error or auc myself, what happens then?</p>\r\n<p>I cannot do it?</p>"
    },
    {
      "id": 24524,
      "postDate": "2013-05-20T15:55:09.827Z",
      "content": "<p>If by results of the test set (test.csv) you mean digits' labels, this is what the competition is all about. We're all trying to figure out what the right digits are. You have to predict labels, submit your guesses as described in submission details, and\r\n you'll obtain score evaluating accuracy of your solution.</p>\r\n<p>&quot;<span>The test data set, (test.csv), is the same as the training set, except that it does not contain the &quot;label&quot; column&quot; - I guess this is the reason of your confusion. It just means that digits in training and test set are described in the same way: pix0,\r\n pix1,...,pix783. The only difference is lack of the labels in test set because you are supposed to predict them.&nbsp;</span></p>"
    },
    {
      "id": 24521,
      "postDate": "2013-05-20T15:11:41.323Z",
      "content": "<p>Hi,</p>\r\n<p>Does anybody know where can I find the actual results of the Test set?</p>\r\n<p>In the Data section it is stated that test set is actually the training set, but when I tried to match training and test set by rows they contain, I don't get any match.</p>\r\n<p>And second the size of training set is different than test set.</p>\r\n<p></p>\r\n<p>I would appreciate if anyone can tell me how can I test my own prediction results.</p>\r\n<p></p>\r\n<p></p>\r\n<p>Best,</p>\r\n<p>Miya</p>\r\n<p></p>"
    }
  ],
  "comments": [
    {
      "id": 24536,
      "author_name": "Filip D",
      "author_url": "",
      "post_date": "2013-05-20T19:49:29.027000",
      "content": "<p>You can split given training data and for instance use 80% of it for training, and 20% for tests.&nbsp;</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 24532,
      "author_name": "miya_clark",
      "author_url": "",
      "post_date": "2013-05-20T19:05:21.087000",
      "content": "<p>Hi,</p>\r\n<p></p>\r\n<p>Thank you for you answer. But If I want to estimate misclassification&nbsp;error or auc myself, what happens then?</p>\r\n<p>I cannot do it?</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 24524,
      "author_name": "Filip D",
      "author_url": "",
      "post_date": "2013-05-20T15:55:09.827000",
      "content": "<p>If by results of the test set (test.csv) you mean digits' labels, this is what the competition is all about. We're all trying to figure out what the right digits are. You have to predict labels, submit your guesses as described in submission details, and\r\n you'll obtain score evaluating accuracy of your solution.</p>\r\n<p>&quot;<span>The test data set, (test.csv), is the same as the training set, except that it does not contain the &quot;label&quot; column&quot; - I guess this is the reason of your confusion. It just means that digits in training and test set are described in the same way: pix0,\r\n pix1,...,pix783. The only difference is lack of the labels in test set because you are supposed to predict them.&nbsp;</span></p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "24536": "",
    "24532": "",
    "24524": "",
    "24521": ""
  }
}