{
  "id": 415338,
  "title": "Hyperparameters Optimization using Optuna",
  "url": "/competitions/predict-student-performance-from-game-play/discussion/415338",
  "author_name": "",
  "post_date": "2023-06-06T04:19:53.037375300Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hello kagglers, this is my first competition and I must admit that I'm learning a lot, I'd like to ask the opinion of experts, I optimized the hyperparameters using kaggle but instead of improving the model it is performing poorer. </p>\n<p>How good is to focus your efforts in tuning the hyperparameters? <br>\nWhat do you recommend most? </p>\n<p><a href=\"https://www.kaggle.com/code/javihm77/catboost-baseline-chapu-optuna#Optuna-Parameters-optimization\" target=\"_blank\">This is my notebook</a></p>",
  "messages": [
    {
      "id": "2289297",
      "postDate": "06/06/2023 04:19:53",
      "content": "<p>Hello kagglers, this is my first competition and I must admit that I'm learning a lot, I'd like to ask the opinion of experts, I optimized the hyperparameters using kaggle but instead of improving the model it is performing poorer. </p>\n<p>How good is to focus your efforts in tuning the hyperparameters? <br>\nWhat do you recommend most? </p>\n<p><a href=\"https://www.kaggle.com/code/javihm77/catboost-baseline-chapu-optuna#Optuna-Parameters-optimization\" target=\"_blank\">This is my notebook</a></p>",
      "rawMarkdown": "Hello kagglers, this is my first competition and I must admit that I'm learning a lot, I'd like to ask the opinion of experts, I optimized the hyperparameters using kaggle but instead of improving the model it is performing poorer. \n\nHow good is to focus your efforts in tuning the hyperparameters? \nWhat do you recommend most? \n\n[This is my notebook](https://www.kaggle.com/code/javihm77/catboost-baseline-chapu-optuna#Optuna-Parameters-optimization)",
      "votes": null
    },
    {
      "id": "2290356",
      "postDate": "06/06/2023 18:04:59",
      "content": "<p>Hyperparameter optimization is still important for traditional ML models (in this competition, I believe most people use the GBDT model), but far less important than the role they play in deep learning models. Few crucial hyperparameters may still give a huge boost to your model, but not all of them. </p>\n<p>According to my experience, feature engineering is the most important part of work if you use a traditional ML model to deal with tabular data.</p>",
      "rawMarkdown": "Hyperparameter optimization is still important for traditional ML models (in this competition, I believe most people use the GBDT model), but far less important than the role they play in deep learning models. Few crucial hyperparameters may still give a huge boost to your model, but not all of them. \n\nAccording to my experience, feature engineering is the most important part of work if you use a traditional ML model to deal with tabular data.",
      "votes": null
    },
    {
      "id": "2290551",
      "postDate": "06/06/2023 23:57:55",
      "content": "<p>Thanks for your answer <a href=\"https://www.kaggle.com/ataraxian\" target=\"_blank\">@ataraxian</a>, it helps me a lot to learn and begin understand what's more important, as I am a beginner I am using tradicional models, xgboost, catboost, random forest, good thing is that I learned to use Optuna 😄</p>",
      "rawMarkdown": "Thanks for your answer @ataraxian, it helps me a lot to learn and begin understand what's more important, as I am a beginner I am using tradicional models, xgboost, catboost, random forest, good thing is that I learned to use Optuna 😄",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2290356,
      "author_name": "ataraxian",
      "author_url": "",
      "post_date": "06/06/2023 18:04:59",
      "content": "<p>Hyperparameter optimization is still important for traditional ML models (in this competition, I believe most people use the GBDT model), but far less important than the role they play in deep learning models. Few crucial hyperparameters may still give a huge boost to your model, but not all of them. </p>\n<p>According to my experience, feature engineering is the most important part of work if you use a traditional ML model to deal with tabular data.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2290551,
          "author_name": "javihm77",
          "author_url": "",
          "post_date": "06/06/2023 23:57:55",
          "content": "<p>Thanks for your answer <a href=\"https://www.kaggle.com/ataraxian\" target=\"_blank\">@ataraxian</a>, it helps me a lot to learn and begin understand what's more important, as I am a beginner I am using tradicional models, xgboost, catboost, random forest, good thing is that I learned to use Optuna 😄</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2289297": "Hello kagglers, this is my first competition and I must admit that I'm learning a lot, I'd like to ask the opinion of experts, I optimized the hyperparameters using kaggle but instead of improving the model it is performing poorer. \n\nHow good is to focus your efforts in tuning the hyperparameters? \nWhat do you recommend most? \n\n[This is my notebook](https://www.kaggle.com/code/javihm77/catboost-baseline-chapu-optuna#Optuna-Parameters-optimization)",
    "2290356": "Hyperparameter optimization is still important for traditional ML models (in this competition, I believe most people use the GBDT model), but far less important than the role they play in deep learning models. Few crucial hyperparameters may still give a huge boost to your model, but not all of them. \n\nAccording to my experience, feature engineering is the most important part of work if you use a traditional ML model to deal with tabular data.",
    "2290551": "Thanks for your answer @ataraxian, it helps me a lot to learn and begin understand what's more important, as I am a beginner I am using tradicional models, xgboost, catboost, random forest, good thing is that I learned to use Optuna 😄"
  },
  "source": "meta"
}