{
  "id": 611248,
  "title": "Can we use some sort of encoder",
  "url": "/competitions/alpha-radar-solana-sprint/discussion/611248",
  "author_name": "",
  "post_date": "2025-10-09T18:04:57.350322300Z",
  "votes": null,
  "comment_count": 7,
  "views": 0,
  "content": "<p>The data is time series and with many columns, can we use an encoder to capture the trend that could be passed in the CatBoost model.<br>\nAsking this as we were told only to use Catboost and no other ML framework</p>",
  "messages": [
    {
      "id": "3300172",
      "postDate": "10/09/2025 18:04:57",
      "content": "<p>The data is time series and with many columns, can we use an encoder to capture the trend that could be passed in the CatBoost model.<br>\nAsking this as we were told only to use Catboost and no other ML framework</p>",
      "rawMarkdown": "The data is time series and with many columns, can we use an encoder to capture the trend that could be passed in the CatBoost model.\nAsking this as we were told only to use Catboost and no other ML framework",
      "votes": null
    },
    {
      "id": "3300173",
      "postDate": "10/09/2025 18:12:35",
      "content": "<p>Sure that's fine. </p>",
      "rawMarkdown": "Sure that's fine.",
      "votes": null
    },
    {
      "id": "3300194",
      "postDate": "10/09/2025 19:45:39",
      "content": "<p>Your submission3, do you have the notebook to verify that?</p>",
      "rawMarkdown": "Your submission3, do you have the notebook to verify that?",
      "votes": null
    },
    {
      "id": "3300269",
      "postDate": "10/10/2025 03:11:23",
      "content": "<p>Ok cool 👍</p>",
      "rawMarkdown": "Ok cool 👍",
      "votes": null
    },
    {
      "id": "3300270",
      "postDate": "10/10/2025 03:12:34",
      "content": "<p>In submission3 I used the sample data only , so i am dropping that , and in that i used the mean , max for trends so it is not really helpful</p>",
      "rawMarkdown": "In submission3 I used the sample data only , so i am dropping that , and in that i used the mean , max for trends so it is not really helpful",
      "votes": null
    },
    {
      "id": "3300272",
      "postDate": "10/10/2025 03:17:34",
      "content": "<p>I'm trying to figure out the best way to not allow the perfect results cause clearly they're not using the spirit of thr competition, just trying to jump up the leaderboard. As far as I'm concerned you're in first place. The .40105 is the best result using the data as it was intended to be used.</p>",
      "rawMarkdown": "I'm trying to figure out the best way to not allow the perfect results cause clearly they're not using the spirit of thr competition, just trying to jump up the leaderboard. As far as I'm concerned you're in first place. The .40105 is the best result using the data as it was intended to be used.",
      "votes": null
    },
    {
      "id": "3300417",
      "postDate": "10/10/2025 11:23:53",
      "content": "<p>Any way to add another dataset for testing, which is completely different from current big dataset? In that way, it would be better to evaluate the model cause' train and test are separated and no one can get 1 because target tokens of the test dataset are not given.</p>",
      "rawMarkdown": "Any way to add another dataset for testing, which is completely different from current big dataset? In that way, it would be better to evaluate the model cause' train and test are separated and no one can get 1 because target tokens of the test dataset are not given.",
      "votes": null
    },
    {
      "id": "3305685",
      "postDate": "10/23/2025 08:41:43",
      "content": "<p>So, I can use the outputs of any model, but it's important that the stacker be a catboost? Isn't it strange that this isn't against the rules considering the restriction is only to catboost?</p>",
      "rawMarkdown": "So, I can use the outputs of any model, but it's important that the stacker be a catboost? Isn't it strange that this isn't against the rules considering the restriction is only to catboost?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3300173,
      "author_name": "twainayar",
      "author_url": "",
      "post_date": "10/09/2025 18:12:35",
      "content": "<p>Sure that's fine. </p>",
      "votes": null,
      "replies": [
        {
          "id": 3300194,
          "author_name": "twainayar",
          "author_url": "",
          "post_date": "10/09/2025 19:45:39",
          "content": "<p>Your submission3, do you have the notebook to verify that?</p>",
          "votes": null,
          "replies": [
            {
              "id": 3300270,
              "author_name": "sakshamm2587",
              "author_url": "",
              "post_date": "10/10/2025 03:12:34",
              "content": "<p>In submission3 I used the sample data only , so i am dropping that , and in that i used the mean , max for trends so it is not really helpful</p>",
              "votes": null,
              "replies": []
            }
          ]
        },
        {
          "id": 3300269,
          "author_name": "sakshamm2587",
          "author_url": "",
          "post_date": "10/10/2025 03:11:23",
          "content": "<p>Ok cool 👍</p>",
          "votes": null,
          "replies": [
            {
              "id": 3300272,
              "author_name": "twainayar",
              "author_url": "",
              "post_date": "10/10/2025 03:17:34",
              "content": "<p>I'm trying to figure out the best way to not allow the perfect results cause clearly they're not using the spirit of thr competition, just trying to jump up the leaderboard. As far as I'm concerned you're in first place. The .40105 is the best result using the data as it was intended to be used.</p>",
              "votes": null,
              "replies": [
                {
                  "id": 3300417,
                  "author_name": "lctngdng",
                  "author_url": "",
                  "post_date": "10/10/2025 11:23:53",
                  "content": "<p>Any way to add another dataset for testing, which is completely different from current big dataset? In that way, it would be better to evaluate the model cause' train and test are separated and no one can get 1 because target tokens of the test dataset are not given.</p>",
                  "votes": null,
                  "replies": []
                }
              ]
            }
          ]
        },
        {
          "id": 3305685,
          "author_name": "darkartitry",
          "author_url": "",
          "post_date": "10/23/2025 08:41:43",
          "content": "<p>So, I can use the outputs of any model, but it's important that the stacker be a catboost? Isn't it strange that this isn't against the rules considering the restriction is only to catboost?</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "3300172": "The data is time series and with many columns, can we use an encoder to capture the trend that could be passed in the CatBoost model.\nAsking this as we were told only to use Catboost and no other ML framework",
    "3300173": "Sure that's fine.",
    "3300194": "Your submission3, do you have the notebook to verify that?",
    "3300269": "Ok cool 👍",
    "3300270": "In submission3 I used the sample data only , so i am dropping that , and in that i used the mean , max for trends so it is not really helpful",
    "3300272": "I'm trying to figure out the best way to not allow the perfect results cause clearly they're not using the spirit of thr competition, just trying to jump up the leaderboard. As far as I'm concerned you're in first place. The .40105 is the best result using the data as it was intended to be used.",
    "3300417": "Any way to add another dataset for testing, which is completely different from current big dataset? In that way, it would be better to evaluate the model cause' train and test are separated and no one can get 1 because target tokens of the test dataset are not given.",
    "3305685": "So, I can use the outputs of any model, but it's important that the stacker be a catboost? Isn't it strange that this isn't against the rules considering the restriction is only to catboost?"
  },
  "source": "meta"
}