{
  "fetched_at": "2026-10-19T19:11:18Z",
  "source_url": "https://www.kaggle.com/competitions/2020fa-inls690-sarcasm-detection/data",
  "pages": [
    {
      "name": "data-description",
      "content": "<h2>File descriptions</h2>\n<ul>\n<li><strong>train.tsv</strong> - Your training data. It has one header line and 53,032 data lines. The first column is the category or label of each training example; the second column is a comment in Reddit; the third column is the parent post. You may or may not find the parent post useful in the classification.</li>\n<li><strong>test.tsv</strong> - Your test data. It has one header line and 17,719 data lines. Same format as the training data, except that the first column is an ID instead of the class label of the line (which your algorithm is going to predict). The ID field in each line in your submission file should match that in the test data.</li>\n<li><strong>sample-submission.csv</strong> - a sample submission file in the correct format. It only contains two columns, id and label. The id's should match the first column of test.tsv exactly, and the second column is your predicted class label (0 or 1) for the corresponding line in text data. Note that the two columns are separated by comma, and please include the exact headers \"id,label\" in your submission.</li>\n</ul>\n<h2>Data fields</h2>\n<ul>\n<li><strong>comment</strong> - [in both train.tsv and test.tsv] the Reddit comment (short text document). It is in free text format. </li>\n<li><strong> parent_comment </strong> [in both train.tsv and test.tsv] the parent post that the comment replies to. It is in free text format. </li>\n<li><strong>label</strong> - [only in train.tsv] the category of a training example. It takes values of either 0 or 1.</li>\n<li><strong>id</strong> - [only in test.tsv] an Id unique to a test example. This number should match to the Id field of the submission file. It should be in the range of 0-17,718.</li>\n</ul>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "<h2>Task Description</h2>\n<p>This is a text classification task. Every document (a line in the data file) contains a Reddit comment and its parent post (the post being commented on), separated with tab.</p>\n\n<p>Your goal is to classify each Reddit comment (not the parent post) into ONE of the two categories, based on whether it is sarcastic or not. The parent post provides necessary context for your judgment.</p>\n\n**0: the comment does NOT express sarcasm**\n\n**1: the comment expresses sarcasm**\n\n<p>The training data contains 53,032 comments, already labeled with one of the above two categories. The test data contains 17,719 comments that are unlabeled. The submission should be a .csv (comma separated free text) file with a header line \"id,label\" followed by exactly 17,719 lines. In each line, there should be exactly two integers, separated by a comma. The first integer is the line ID of a test example (0, 1, 2, ..., 17718), and the second integer is the category (or label) predicted by your classifier, one of {0,1}.</p>\n\n<p>You can make 10 submissions per day. Once you submit your results, you will get an accuracy score computed based on 50% of the test data. This score will position you somewhere on the leaderboard. Once the competition ends, you will see the final accuracy computed based on the other 50% of the test data. The evaluation metric is the accuracy of your classifier. It is calculated as (# of correctly classified test examples)/(total # of test examples), so the higher the better.</p>\n\n<p>You can use any classifier or combination of classifiers, any combination or selection of features, and either supervised, semi-supervised, or even transfer learning approaches. You can also be creative and make use of external data sources that do not contain the exact comments or parent posts in Reddit, or the original larger data set by Dan Ofer.</p>\n\n<h2>Acknowledgements</h2>\n<p>We thank Dan Ofer and Professor Plum for providing the data.</p>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "<p>The evaluation metric for this competition is accuracy - the higher the better. </p>\n\n\nSubmission Format\n-----------------\n\n\n<p>For each line in the test data, submission files should contain two fields: id and label. Every line should contain two integers, separated by a comma. The first integer tells the line ID of a test question (0 to 17,718), and the second integer is your prediction of its category (0 or 1). </p>\n\n<p>The file should contain a header and have the following format:</p>\n\n<pre>id,label<br />0,1<br />1,1<br />2,0<br />3,0<br />...<br />17718,0</pre>\n<p>Do not include the original post content in your submission file!!</p>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": " - Do not link back to the exact same comments or parent posts in Reddit, or the original data set.\n - Have fun!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}