{
  "fetched_at": "2026-10-20T08:56:29Z",
  "source_url": "https://www.kaggle.com/competitions/detect-skills-from-text/data",
  "pages": [
    {
      "name": "data-description",
      "content": "# Files\n\n## skills.csv \nThis file represents the unique skills that you should try to link to texts. The file has 3 columns:\n\n* ``skill_id`` - the unique identifier of the skill (should be used in the list of skills that you provide as the result variable).\n* ``skill_name`` - The name of the skill (in Estonian).\n* `skill_description` - The description of the skill (in Estonian).\n\n## texts_about_skills.csv\nThis file includes the raw texts that need to be mapped with the skills. The file has 2 columns:\n* `text_id` - the unique identifier of the text (these identifiers are used both in train and test files).\n* `text` - the raw text from where you should find the skills (in Estonian).\n\n## train.csv \nTraining set that should be used to build your machine learning models. The training set includes both the text identificator and the ground truth about the ``skills`` that correspond to the text:\n* ``text_id`` – the id of the text,\n* ``skills`` – space separated list of skill ids that correspond to this text.\n\n## test.csv \nTest set should be used to see how well your model performs on unseen data. The test set includes the ``text_id`` but not the ``skills`` that correspond to the text. It is your job to predict the skills using the model you built on the training data!\n\n## sample_submission.csv\nAn example of what the submission file should look like. For every row in the test data, submission files should contain two columns: ``text_id`` (representing the id of the text) and ``skills`` (a **space separated list of the skills** corresponding to the text). Example format:\n\n``text_id, skills``\n\n``text0, s0``\n\n``text108, s11 s22 s31 s41``\n\n``Etc.``\n\n\nIn the example above, we say that text with id **text108** includes four skills **s11 s22 s31 and s41**. All unique skills are described in ``skills.csv`` (read more under data section).\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "Estonian government AI testbed competition invites everyone with expertise in data science and analysis to join an experimental testbed to find solutions to real-world problems using open data. We look forward to welcoming a range of specialists to come and work with us and become part of one of the most technologically advanced government projects in Europe contributing to the development of AI Gov stack that benefits people across the globe. The AI testbed competition encompasses various works in the fields of data science, machine learning, and language technology, that open a way to contribute and become a part of developing the open-source AI Govstack. \n\nPick a competition you are interested in and help us to build the coolEST and brightEST digital state!\n\n# Background\nHave you ever hesitated whilst answering the question “Who do you want to become?”. Statistics show that less than 10% of the people know exactly what career path they would like to choose. However not only workers struggle with finding the right position, but HR specialists as well face difficulties when searching for a perfect match. One of the keys for success in choosing the right candidate is having their skills match the task they are going to fulfill. This process can be quite time-consuming and unfeasible for the HR specialists due to ever changing nature of the underlying documents (professional standards, curricula, syllabuses, surveys, job advertisements, websites).  \n\nThe accurate skill detection model would result in big time saving for analysts, school principals, personnel specialists, career counselors and for common person planning their own learning and career path.\n\n# Goal\nThe goal of this competition is to detect skills that correspond to the texts about forestry/wood industry and metal/machine industry. \n\n# Praise\n* The best solutions will be publicly introduced to Estonian public sector.\n* The best solutions become a part of Estonian e-government [code repository]( https://koodivaramu.eesti.ee/explore/projects).\n* Social impact by creating a solution which would benefit a society as a whole.\n* Chance to virtually meet top Estonian data scientists and build your network.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "# Metric\nThe evaluation metric for this competition is [Mean F-Score](https://en.wikipedia.org/wiki/F-score) or the arithmetic mean of individual F-scores. The F-score weights recall and precision equally. \n\n*-Precision is the ratio of true positives to all predicted positives.*\n*-Recall is the ratio of true positives to all actual positives.* \n\n# Submission format\nYou should submit a csv file with the same number of rows + header row as in test.csv file. For every row in the test data, submission files should contain two columns: ``text_id`` (representing the id of the text) and ``skills`` (a **space separated list of the skills** corresponding to the text). Example format:\n```\ntext_id, skills\ntext0, s0\ntext108, s11 s22 s31 s41\nEtc.\n```\n\nIn the example above, we say that text with id **text108** includes four skills **s11 s22 s31 and s41**. All unique skills are described in ``skills.csv`` (read more under data section). \n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "# Open code\nThis competition is all about openness and you should be willing to publish your code. The easiest way to do this is to publish your solution as a Kaggle notebook. This is easily done by going to the “Code” tab within a competition’s page and then clicking “New Notebook.” \n\n# Team limits\nThere is no minimum limit to the team size (you can work alone if you wish to do so).\n\n# Submission limits\n* You may submit a maximum of 5 entries per day.\n* You may select up to 3 final submissions for judging.\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}