{
  "id": 71072,
  "title": "How is the Leaderboard Scored",
  "url": "/competitions/ga-customer-revenue-prediction/discussion/71072",
  "author_name": "",
  "post_date": "2018-11-09T21:54:21.095546Z",
  "votes": null,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Does anyone know how is the LB is scored? My understanding is, if I am right, we need to predict customer in test_v2. If so, why our previous scores are still there, since the old submission has over 600k row, test_v2 only 297k rows. Furthermore, do we need to build a new model by using train_v2.csv and to predict test_2v? I am very confused at this moment.</p>",
  "messages": [
    {
      "id": "418415",
      "postDate": "11/09/2018 21:54:21",
      "content": "<p>Does anyone know how is the LB is scored? My understanding is, if I am right, we need to predict customer in test_v2. If so, why our previous scores are still there, since the old submission has over 600k row, test_v2 only 297k rows. Furthermore, do we need to build a new model by using train_v2.csv and to predict test_2v? I am very confused at this moment.</p>",
      "rawMarkdown": "Does anyone know how is the LB is scored? My understanding is, if I am right, we need to predict customer in test_v2. If so, why our previous scores are still there, since the old submission has over 600k row, test_v2 only 297k rows. Furthermore, do we need to build a new model by using train_v2.csv and to predict test_2v? I am very confused at this moment.",
      "votes": null
    },
    {
      "id": "418422",
      "postDate": "11/09/2018 22:05:15",
      "content": "<p>I agree the leaderboard needs to be cleared, this is just messy.</p>",
      "rawMarkdown": "I agree the leaderboard needs to be cleared, this is just messy.",
      "votes": null
    },
    {
      "id": "418434",
      "postDate": "11/09/2018 22:17:57",
      "content": "<p>Exactly, this is very annoying. They should clean the LB</p>",
      "rawMarkdown": "Exactly, this is very annoying. They should clean the LB",
      "votes": null
    },
    {
      "id": "418436",
      "postDate": "11/09/2018 22:27:20",
      "content": "<p>Even if they clean it, the test_v2 = Leaderboard..</p>\n\n<p>You can see that here : <a href=\"https://www.kaggle.com/paulorzp/1line-perfect-score\">https://www.kaggle.com/paulorzp/1line-perfect-score</a></p>",
      "rawMarkdown": "Even if they clean it, the test_v2 = Leaderboard..\n\nYou can see that here : https://www.kaggle.com/paulorzp/1line-perfect-score",
      "votes": null
    },
    {
      "id": "418439",
      "postDate": "11/09/2018 22:38:09",
      "content": "<p>So, do we need to use the train_v2 to build a new model and predict on test_v2 only?</p>",
      "rawMarkdown": "So, do we need to use the train_v2 to build a new model and predict on test_v2 only?",
      "votes": null
    },
    {
      "id": "418442",
      "postDate": "11/09/2018 22:42:34",
      "content": "<p>Exactly :) \nThis is why your cross-validation strategy will be very important in this competition !</p>",
      "rawMarkdown": "Exactly :) \nThis is why your cross-validation strategy will be very important in this competition !",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 418422,
      "author_name": "dmitriyguller",
      "author_url": "",
      "post_date": "11/09/2018 22:05:15",
      "content": "<p>I agree the leaderboard needs to be cleared, this is just messy.</p>",
      "votes": null,
      "replies": [
        {
          "id": 418434,
          "author_name": "lftuwujie",
          "author_url": "",
          "post_date": "11/09/2018 22:17:57",
          "content": "<p>Exactly, this is very annoying. They should clean the LB</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 418436,
          "author_name": "xtelima",
          "author_url": "",
          "post_date": "11/09/2018 22:27:20",
          "content": "<p>Even if they clean it, the test_v2 = Leaderboard..</p>\n\n<p>You can see that here : <a href=\"https://www.kaggle.com/paulorzp/1line-perfect-score\">https://www.kaggle.com/paulorzp/1line-perfect-score</a></p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 418439,
          "author_name": "lftuwujie",
          "author_url": "",
          "post_date": "11/09/2018 22:38:09",
          "content": "<p>So, do we need to use the train_v2 to build a new model and predict on test_v2 only?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 418442,
          "author_name": "xtelima",
          "author_url": "",
          "post_date": "11/09/2018 22:42:34",
          "content": "<p>Exactly :) \nThis is why your cross-validation strategy will be very important in this competition !</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "418415": "Does anyone know how is the LB is scored? My understanding is, if I am right, we need to predict customer in test_v2. If so, why our previous scores are still there, since the old submission has over 600k row, test_v2 only 297k rows. Furthermore, do we need to build a new model by using train_v2.csv and to predict test_2v? I am very confused at this moment.",
    "418422": "I agree the leaderboard needs to be cleared, this is just messy.",
    "418434": "Exactly, this is very annoying. They should clean the LB",
    "418436": "Even if they clean it, the test_v2 = Leaderboard..\n\nYou can see that here : https://www.kaggle.com/paulorzp/1line-perfect-score",
    "418439": "So, do we need to use the train_v2 to build a new model and predict on test_v2 only?",
    "418442": "Exactly :) \nThis is why your cross-validation strategy will be very important in this competition !"
  },
  "source": "meta"
}