{
  "id": 42446,
  "title": "Query on timeseries modelling and Na values",
  "url": "/competitions/favorita-grocery-sales-forecasting/discussion/42446",
  "author_name": "Nagaraju Oruganti",
  "post_date": "2017-10-31T04:22:44.337000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I am planing to predict sales of each item based on their sales in the previous year.I considered only August data for timebeing My dataset looks like this (attachement :sec-2)</p>\n\n<p>I am planing to train the model on 2013 - 2016 dataset and predict on 2017 (attachement:sec-1) </p>\n\n<p>I am planing to sub missing values with 0, since missing values represent no sales during that period (attachement: sec-3) </p>\n\n<p>I am bit confused how to treat missing values in this case. I think this product has been launched in 2015 and hence 2013 and 2014 values are missing. Replacing with 0 may not be correct way of dealing with Na values here(attachement:sec-4) </p>\n\n<p>Same scenario as above ,this product has been launched in 2016. Pls help me in missing value treatment (attachement:sec-5) </p>\n\n<p>This product has been launched in 2017.Hence no trainingset value available for August. Kindly assist with the process of forecasting Anticipating useful advice/assistance in this regard. Thanks in advance. Let me know if my explanation is clear or not. Will elaborate if needed</p>",
  "messages": [
    {
      "id": 237843,
      "postDate": "2017-10-31T04:22:44.337Z",
      "content": "<p>I am planing to predict sales of each item based on their sales in the previous year.I considered only August data for timebeing My dataset looks like this (attachement :sec-2)</p>\n\n<p>I am planing to train the model on 2013 - 2016 dataset and predict on 2017 (attachement:sec-1) </p>\n\n<p>I am planing to sub missing values with 0, since missing values represent no sales during that period (attachement: sec-3) </p>\n\n<p>I am bit confused how to treat missing values in this case. I think this product has been launched in 2015 and hence 2013 and 2014 values are missing. Replacing with 0 may not be correct way of dealing with Na values here(attachement:sec-4) </p>\n\n<p>Same scenario as above ,this product has been launched in 2016. Pls help me in missing value treatment (attachement:sec-5) </p>\n\n<p>This product has been launched in 2017.Hence no trainingset value available for August. Kindly assist with the process of forecasting Anticipating useful advice/assistance in this regard. Thanks in advance. Let me know if my explanation is clear or not. Will elaborate if needed</p>",
      "rawMarkdown": "I am planing to predict sales of each item based on their sales in the previous year.I considered only August data for timebeing My dataset looks like this (attachement :sec-2)\n\n \nI am planing to train the model on 2013 - 2016 dataset and predict on 2017 (attachement:sec-1) \n\nI am planing to sub missing values with 0, since missing values represent no sales during that period (attachement: sec-3) \n\nI am bit confused how to treat missing values in this case. I think this product has been launched in 2015 and hence 2013 and 2014 values are missing. Replacing with 0 may not be correct way of dealing with Na values here(attachement:sec-4) \n\nSame scenario as above ,this product has been launched in 2016. Pls help me in missing value treatment (attachement:sec-5) \n\nThis product has been launched in 2017.Hence no trainingset value available for August. Kindly assist with the process of forecasting Anticipating useful advice/assistance in this regard. Thanks in advance. Let me know if my explanation is clear or not. Will elaborate if needed\n\n"
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "237843": "I am planing to predict sales of each item based on their sales in the previous year.I considered only August data for timebeing My dataset looks like this (attachement :sec-2)\n\n \nI am planing to train the model on 2013 - 2016 dataset and predict on 2017 (attachement:sec-1) \n\nI am planing to sub missing values with 0, since missing values represent no sales during that period (attachement: sec-3) \n\nI am bit confused how to treat missing values in this case. I think this product has been launched in 2015 and hence 2013 and 2014 values are missing. Replacing with 0 may not be correct way of dealing with Na values here(attachement:sec-4) \n\nSame scenario as above ,this product has been launched in 2016. Pls help me in missing value treatment (attachement:sec-5) \n\nThis product has been launched in 2017.Hence no trainingset value available for August. Kindly assist with the process of forecasting Anticipating useful advice/assistance in this regard. Thanks in advance. Let me know if my explanation is clear or not. Will elaborate if needed\n\n"
  }
}