{
  "fetched_at": "2026-10-19T15:31:42Z",
  "source_url": "https://www.kaggle.com/competitions/dl-course-final-competition/data",
  "pages": [
    {
      "name": "data-description",
      "content": "# File description\n* **train_data.csv** - the training set for your model training\n* **test_data.csv** - the testing set for the leaderboard ranking and final scoring\n* **sample.csv** - the sample submission file in the correct format\n\n# Data fields\n* **id** - an anonymous id unique to a given news title\n* **label** - the label category of the news title\n* **label_name** - the category name of the news title\n* **title** - the title of the news\n* **keyword** - the keywords from the news",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "# Introduction\nThis is the final competition for NCTU Deep Learning Course.\nIn this competition, you need to do the classification problem for the Chinese news title in different categories. \n\nThere has 10 different categories for the Chinese news title. You need to use the model to predict the category based on the news title or some keywords.\n\n* 娛樂 news_entertainment\n* 體育 news_sports\n* 房產 news_house\n* 汽車 news_car\n* 教育 news_edu \n* 科技 news_tech\n* 軍事 news_military\n* 國際 news_world\n* 農業 news_agriculture\n* 電競 news_game\n\n\n# Upload\n1. You need to upload the **answer.csv** file for your result of prediction\n2. The detail of the format, please check the [Evaluation](https://www.kaggle.com/c/dl-course-final-competition/overview/evaluation) and [Data](https://www.kaggle.com/c/dl-course-final-competition/data)\n3. You need to upload your code and README to New E3 system\n\n# Rules\n1. Only one person as a team\n2. The team name should name as your student ID, i.e. 08XXXXX, AXXXXXX, BXXXXXX\n3. You can submit the answer **20** times every 24 hours\n4. You can reference the code and architecture for your model on the internet\n5. Do not share your result with your classmate\n6. The score will base on the result of the private leaderboard\n\n# Hint\nYou should tokenize the Chinese word to create the dictionary\n* [Jieba](https://github.com/fxsjy/jieba)\n* [CKIP](https://github.com/ldkrsi/ckip_python)\n* [NLTK](https://www.nltk.org/)\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "The evaluation metric for this competition is [Accuracy](https://en.wikipedia.org/wiki/Accuracy_and_precision).\nThe Accuracy = (TP + TN)/(TP + TN + FP + FN)\n# Submission Format\nSubmission files should contain two columns: id and label ( label should be a [integer](https://numpy.org/devdocs/user/basics.types.html) between 0 and 9)\n\n<p>The file should contain a header and have the following format:</p>\n<pre>id,label<br>0,2<br>1,1<br>   .<br>   .<br>   .<br>59907,5<br></pre>\n\nYou can use any name but with file type **.csv**\nFor more details, you can reference the [sample.csv](https://www.kaggle.com/c/dl-course-final-competition/data)",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "# Rules\n1. Only one person as a team\n2. The team name should name as your student ID, i.e. 08XXXXX, AXXXXXX, BXXXXXX\n3. You can submit the answer **20** times every 24 hours\n4. You can reference the code and architecture for your model on the internet\n5. Do not share your result with your classmate\n6. The score will be based on the result of the private leaderboard",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}