{
  "fetched_at": "2026-10-19T07:43:00Z",
  "source_url": "https://www.kaggle.com/competitions/2021mlfinal/data",
  "pages": [
    {
      "name": "data-description",
      "content": "# Dataset\n\n## Train/Valid/Test data: 2012-01-01 ~ 2014-12-31 (3 years) / 2015-01-01 ~ 2015-12-31 (1 year) / 2016-01-01 ~ 2016-12-31 (1 year) for 141 stocks with market index, respectively.\n\n*   Input: 11 numeric features of 60 previous days for each stocks, the index 0 is for the market index. (shape of x_t: [142 (1 + 141), 60, 11], Reference: Enhancing Stock Movement Prediction with Adversarial Training (Feng et al., IJCAI-19))\n\n*   Target: Closing price ratio over the previous day for a given date of each stocks (shape: [141])",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "# Fundamentals of Machine Learning (Fall 2021)\n\n## Final Project (Stock Closing Price Prediction)\n\n- Build a machine learning model to predict the closing price of KOSPI stocks.\n- Kaggle & Colab Guide: Kaggle & Colab Manual 2021.pptx\n- Tutorials:\n(1) Basics: https://bit.ly/pytorch_1_basics\n(2) Modules: https://bit.ly/pytorch_2_modules\n(3) Building a Model: https://bit.ly/pytorch_3_building_model\n(4) Train and Predict: https://bit.ly/pytorch_4_train_predict\n- Write your answers on ‘Answer.ipynb’: https://bit.ly/2021_ml_final\n- Basline code:\n(1) DecisionTreeRegressor: https://bit.ly/final_basline_dt\n(2) Recurrent Neural Network: https://bit.ly/final_baseline_rnn\n- Reference:\nhttps://github.com/Ceruleanacg/Personae\nhttps://github.com/LastAncientOne/Deep-Learning-Machine-Learning-Stock\n- Due date: 2021/12/15 23:59:59 KST",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "The evaluation metric for this competition is [Mean Absolute Error](https://en.wikipedia.org/wiki/Mean_absolute_error). Mean absolute error (MAE) is a measure of errors between paired observations expressing the same phenomenon.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "1.\tYou should only train your model with the given data. Any use of external data is prohibited.\n2.\tYou are allowed to write your code in ‘Answer.ipynb’ only.\n3.\tLate submission on either Kaggle or report will not be accepted.\n4.\tYour submission to Kaggle is limited to 10 times a day.\n5.\tOnly specified libraries are permitted to use.\n- Python basic library, Numpy, Pandas, Scikit-learn, Matplotlib, Pytorch, Tensorflow, Keras + dependent packages installed by them.\n6.\tAny kind of action or code considered as cheating is strongly prohibited.\n7.      You must predict the closing price ratio r_t with only subset of the given time-series input data [x_(t-60),…,x_(t-1)]. Using other data not mentioned is considered as cheating.\n\nWe will provide the accuracy of baseline models (DecisionTreeRegressor, RNN) for you to check your performance.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}