{
  "id": 21670,
  "title": "Data censored not at random",
  "url": "/competitions/grupo-bimbo-inventory-demand/discussion/21670",
  "author_name": "Jianbo Ye",
  "post_date": "2016-06-14T08:36:25.997000",
  "votes": 3,
  "comment_count": 0,
  "views": 626,
  "content": "<p>I am a bit confused about the goal of this competition, although I am enjoying the fun to play with such data. Obviously, products are not delivered every month, the records in the dataset is censored not at random: that said, for each row, at least one of Venta_uni_hoy, Dev_uni_proxima is not zero. </p>\n\n<p>It makes me to think that the goal of this competition is NOT to predict the &quot;real&quot; demand of products, but rather to (conditionally) predict Demanda_uni_equil if there has been observed delivery event. A delivery event indicates a row in train/test data. However, one can not know the delivery call beforehand, unless the retailer decides to import and return certain amount of products. As such, the Venta_uni_hoy/Dev_uni_proxima often comes with the delivery call. </p>\n\n<p>Is this a realistic challenge?</p>",
  "messages": [
    {
      "id": 123866,
      "postDate": "2016-06-14T08:36:25.997Z",
      "content": "<p>I am a bit confused about the goal of this competition, although I am enjoying the fun to play with such data. Obviously, products are not delivered every month, the records in the dataset is censored not at random: that said, for each row, at least one of Venta_uni_hoy, Dev_uni_proxima is not zero. </p>\n\n<p>It makes me to think that the goal of this competition is NOT to predict the &quot;real&quot; demand of products, but rather to (conditionally) predict Demanda_uni_equil if there has been observed delivery event. A delivery event indicates a row in train/test data. However, one can not know the delivery call beforehand, unless the retailer decides to import and return certain amount of products. As such, the Venta_uni_hoy/Dev_uni_proxima often comes with the delivery call. </p>\n\n<p>Is this a realistic challenge?</p>",
      "rawMarkdown": "I am a bit confused about the goal of this competition, although I am enjoying the fun to play with such data. Obviously, products are not delivered every month, the records in the dataset is censored not at random: that said, for each row, at least one of Venta_uni_hoy, Dev_uni_proxima is not zero. \r\n\r\nIt makes me to think that the goal of this competition is NOT to predict the \"real\" demand of products, but rather to (conditionally) predict Demanda_uni_equil if there has been observed delivery event. A delivery event indicates a row in train/test data. However, one can not know the delivery call beforehand, unless the retailer decides to import and return certain amount of products. As such, the Venta_uni_hoy/Dev_uni_proxima often comes with the delivery call. \r\n\r\nIs this a realistic challenge?",
      "votes": 3
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "123866": "I am a bit confused about the goal of this competition, although I am enjoying the fun to play with such data. Obviously, products are not delivered every month, the records in the dataset is censored not at random: that said, for each row, at least one of Venta_uni_hoy, Dev_uni_proxima is not zero. \r\n\r\nIt makes me to think that the goal of this competition is NOT to predict the \"real\" demand of products, but rather to (conditionally) predict Demanda_uni_equil if there has been observed delivery event. A delivery event indicates a row in train/test data. However, one can not know the delivery call beforehand, unless the retailer decides to import and return certain amount of products. As such, the Venta_uni_hoy/Dev_uni_proxima often comes with the delivery call. \r\n\r\nIs this a realistic challenge?"
  }
}