{
  "id": 71449,
  "title": "How to make predictions ?",
  "url": "/competitions/ga-customer-revenue-prediction/discussion/71449",
  "author_name": "",
  "post_date": "2018-11-13T19:48:15.497030200Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Really basic question </p>\n\n<p>We have to predict for all the users in sample submission v2.csv. But we dont have any information like geographical information,hits etc.</p>\n\n<p>So how are you gonna predict the target variable ?</p>",
  "messages": [
    {
      "id": "420546",
      "postDate": "11/13/2018 19:48:15",
      "content": "<p>Really basic question </p>\n\n<p>We have to predict for all the users in sample submission v2.csv. But we dont have any information like geographical information,hits etc.</p>\n\n<p>So how are you gonna predict the target variable ?</p>",
      "rawMarkdown": "Really basic question \n\nWe have to predict for all the users in sample submission v2.csv. But we dont have any information like geographical information,hits etc.\n\nSo how are you gonna predict the target variable ?",
      "votes": null
    },
    {
      "id": "420581",
      "postDate": "11/13/2018 21:01:37",
      "content": "<p>I think the point is to consider features like geographical info, hits and even the revenue from the period May 1st to Oct 15th.  Based on these input data, you have to predict the future revenue of each user (fullVisitorIds that entered the web in the aforementioned period). </p>\n\n<p>You simply can't know information about how many sessions each user will have, how many hits per session, etc. Your inputs have to be based on the period May 1st to Oct 15th.</p>\n\n<p>Hope this helps :)</p>",
      "rawMarkdown": "I think the point is to consider features like geographical info, hits and even the revenue from the period May 1st to Oct 15th.  Based on these input data, you have to predict the future revenue of each user (fullVisitorIds that entered the web in the aforementioned period). \n\nYou simply can't know information about how many sessions each user will have, how many hits per session, etc. Your inputs have to be based on the period May 1st to Oct 15th.\n\nHope this helps :)",
      "votes": null
    },
    {
      "id": "421174",
      "postDate": "11/14/2018 17:06:56",
      "content": "<p>So, Can we think the problem as a Timeseries Forecasting Problem for fullVistitorId ?</p>",
      "rawMarkdown": "So, Can we think the problem as a Timeseries Forecasting Problem for fullVistitorId ?",
      "votes": null
    },
    {
      "id": "421615",
      "postDate": "11/15/2018 07:37:28",
      "content": "<p>I think Niels' answer is very good (<a href=\"https://www.kaggle.com/c/ga-customer-revenue-prediction/discussion/71547#421197\">https://www.kaggle.com/c/ga-customer-revenue-prediction/discussion/71547#421197</a>)</p>",
      "rawMarkdown": "I think Niels' answer is very good (https://www.kaggle.com/c/ga-customer-revenue-prediction/discussion/71547#421197)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 420581,
      "author_name": "sgb1710",
      "author_url": "",
      "post_date": "11/13/2018 21:01:37",
      "content": "<p>I think the point is to consider features like geographical info, hits and even the revenue from the period May 1st to Oct 15th.  Based on these input data, you have to predict the future revenue of each user (fullVisitorIds that entered the web in the aforementioned period). </p>\n\n<p>You simply can't know information about how many sessions each user will have, how many hits per session, etc. Your inputs have to be based on the period May 1st to Oct 15th.</p>\n\n<p>Hope this helps :)</p>",
      "votes": null,
      "replies": [
        {
          "id": 421174,
          "author_name": "praneethkumargv",
          "author_url": "",
          "post_date": "11/14/2018 17:06:56",
          "content": "<p>So, Can we think the problem as a Timeseries Forecasting Problem for fullVistitorId ?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 421615,
          "author_name": "sgb1710",
          "author_url": "",
          "post_date": "11/15/2018 07:37:28",
          "content": "<p>I think Niels' answer is very good (<a href=\"https://www.kaggle.com/c/ga-customer-revenue-prediction/discussion/71547#421197\">https://www.kaggle.com/c/ga-customer-revenue-prediction/discussion/71547#421197</a>)</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "420546": "Really basic question \n\nWe have to predict for all the users in sample submission v2.csv. But we dont have any information like geographical information,hits etc.\n\nSo how are you gonna predict the target variable ?",
    "420581": "I think the point is to consider features like geographical info, hits and even the revenue from the period May 1st to Oct 15th.  Based on these input data, you have to predict the future revenue of each user (fullVisitorIds that entered the web in the aforementioned period). \n\nYou simply can't know information about how many sessions each user will have, how many hits per session, etc. Your inputs have to be based on the period May 1st to Oct 15th.\n\nHope this helps :)",
    "421174": "So, Can we think the problem as a Timeseries Forecasting Problem for fullVistitorId ?",
    "421615": "I think Niels' answer is very good (https://www.kaggle.com/c/ga-customer-revenue-prediction/discussion/71547#421197)"
  },
  "source": "meta"
}