{
  "fetched_at": "2026-10-20T08:06:05Z",
  "source_url": "https://www.kaggle.com/competitions/307matrixcompletion2019/data",
  "pages": [
    {
      "name": "rules",
      "content": "This is a page where you can include rules that participants must accept before joining. You may wish to include rules like:\n\n - Don't cheat!\n - Apply yourself!\n - Have fun!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "data-description",
      "content": "<p>Just one file, which contains a low-rank matrix with missing entries.  Make sure your process is using the included Matlab mfiles.</p>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "<p>The data consists of a subset of the entries of a low-rank matrix. &nbsp;Your goal is to fill in the entries marked NaN with numerical values. &nbsp;NaN stands for \"Not-a-Number\". &nbsp;Type &gt;&gt; help NaN in Matlab to learn how to work with NaN values.</p>\n<p>In order to take the Kaggle data and put it into matrix form, use the Matlab program csvReadFromKaggle.m. &nbsp;It is included in this kaggle as part of the \"data\". &nbsp;It takes as input the name of the csv file.</p>\n<p>Example: &nbsp;M =csvReadFromKaggle('Mobserved50by200').</p>\n<p>The matrix M now consists of numerical values for observed entries and NaN's for unobserved entries. &nbsp;Now (somehow) make a matrix Mhat which has as entry (i,j) your estimate of the corresponding entry the original matrix.</p>\n<p>In order to upload your estimate to Kaggle, use the Matlab function csvWriteForKaggle. &nbsp;Upload the resulting .csv file to Kaggle.</p>\n<p>Example: &nbsp;csvWriteForKaggle('mySubmission', Mhat)</p>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "<p>Let Mhat be the matrix in your submission and M be the original low-rank matrix. &nbsp;The score is the Hilbert-Schmidt norm of the difference, multiplied by a certain scalar. &nbsp;A low score is better. &nbsp;A score of 0 would mean perfect matrix completion, but that's not possible, is it?</p>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}