{
  "id": "2002.02046",
  "version": "v1",
  "published": "2020-02-06T00:57:39.000Z",
  "updated": "2020-02-06T00:57:39.000Z",
  "title": "Supervised Learning on Relational Databases with Graph Neural Networks",
  "authors": [
    "Milan Cvitkovic"
  ],
  "categories": [
    "cs.LG",
    "cs.AI",
    "cs.DB",
    "stat.ML"
  ],
  "abstract": "The majority of data scientists and machine learning practitioners use relational data in their work [State of ML and Data Science 2017, Kaggle, Inc.]. But training machine learning models on data stored in relational databases requires significant data extraction and feature engineering efforts. These efforts are not only costly, but they also destroy potentially important relational structure in the data. We introduce a method that uses Graph Neural Networks to overcome these challenges. Our proposed method outperforms state-of-the-art automatic feature engineering methods on two out of three datasets.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2020-02-06T00:57:39.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "graph neural networks",
      "relational databases",
      "potentially important relational structure",
      "automatic feature engineering methods",
      "state-of-the-art automatic feature engineering"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}