{
  "id": 459616,
  "title": "39th - Implementation of PrecTime Model from Arxiv Paper in This Competition.",
  "url": "/competitions/child-mind-institute-detect-sleep-states/discussion/459616",
  "author_name": "Zhecheng Li",
  "post_date": "2023-12-06T02:19:52.548000",
  "votes": 14,
  "comment_count": 0,
  "views": 0,
  "content": "<p><strong>In this competition, we utilized the PrecTime model, which had satisfactory results. Hence, we are sharing it here with everyone.</strong></p>\n<p><strong>As the original paper did not provide open-source code, we attempted to reproduce it. However, there might be some misunderstanding of its structure, which could have prevented us from reaching its highest potential score. I believe other teams might have also used this model. Looking forward to everyone's discussion.</strong></p>\n<p>Paper: <a href=\"https://arxiv.org/abs/2302.10182\" target=\"_blank\">PrecTime: A Deep Learning Architecture for Precise Time Series Segmentation in Industrial Manufacturing Operations</a><br>\nOur Code: <a href=\"https://github.com/Lizhecheng02/Kaggle-Detect_Sleep_States\" target=\"_blank\">Here</a></p>\n<p>Main Structure:<br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F10671485%2F854eb1755cd5e1a63104e4f4cc0ac248%2F_20231205181014.png?generation=1701828638336931&amp;alt=media\" alt=\"\"></p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F10671485%2F2f25d9b761b92ffefabad60dba87b843%2F_20231205181019.png?generation=1701828815897346&amp;alt=media\" alt=\"\"></p>\n<pre><code>:\n.kfold0: . kfold1: . kfold2: . kfold3: . kfold4: . avg: . lb: .\n.kfold0: . kfold1: . kfold2: . kfold3: . kfold4: . avg: . lb: .\n</code></pre>\n<p><strong>Finally, congratulations to all the medal winners, now you can have a good sleep.</strong></p>",
  "messages": [
    {
      "id": 2550382,
      "postDate": "2023-12-06T02:19:52.550Z",
      "content": "<p><strong>In this competition, we utilized the PrecTime model, which had satisfactory results. Hence, we are sharing it here with everyone.</strong></p>\n<p><strong>As the original paper did not provide open-source code, we attempted to reproduce it. However, there might be some misunderstanding of its structure, which could have prevented us from reaching its highest potential score. I believe other teams might have also used this model. Looking forward to everyone's discussion.</strong></p>\n<p>Paper: <a href=\"https://arxiv.org/abs/2302.10182\" target=\"_blank\">PrecTime: A Deep Learning Architecture for Precise Time Series Segmentation in Industrial Manufacturing Operations</a><br>\nOur Code: <a href=\"https://github.com/Lizhecheng02/Kaggle-Detect_Sleep_States\" target=\"_blank\">Here</a></p>\n<p>Main Structure:<br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F10671485%2F854eb1755cd5e1a63104e4f4cc0ac248%2F_20231205181014.png?generation=1701828638336931&amp;alt=media\" alt=\"\"></p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F10671485%2F2f25d9b761b92ffefabad60dba87b843%2F_20231205181019.png?generation=1701828815897346&amp;alt=media\" alt=\"\"></p>\n<pre><code>:\n.kfold0: . kfold1: . kfold2: . kfold3: . kfold4: . avg: . lb: .\n.kfold0: . kfold1: . kfold2: . kfold3: . kfold4: . avg: . lb: .\n</code></pre>\n<p><strong>Finally, congratulations to all the medal winners, now you can have a good sleep.</strong></p>",
      "rawMarkdown": "**In this competition, we utilized the PrecTime model, which had satisfactory results. Hence, we are sharing it here with everyone.**\n\n**As the original paper did not provide open-source code, we attempted to reproduce it. However, there might be some misunderstanding of its structure, which could have prevented us from reaching its highest potential score. I believe other teams might have also used this model. Looking forward to everyone's discussion.**\n\nPaper: [PrecTime: A Deep Learning Architecture for Precise Time Series Segmentation in Industrial Manufacturing Operations](https://arxiv.org/abs/2302.10182)\nOur Code: [Here](https://github.com/Lizhecheng02/Kaggle-Detect_Sleep_States)\n\nMain Structure:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F10671485%2F854eb1755cd5e1a63104e4f4cc0ac248%2F_20231205181014.png?generation=1701828638336931&alt=media)\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F10671485%2F2f25d9b761b92ffefabad60dba87b843%2F_20231205181019.png?generation=1701828815897346&alt=media)\n\n```\nResults:\n1.kfold0: 0.743 kfold1: 0.769 kfold2: 0.793 kfold3: 0.761 kfold4: 0.772 avg: 0.7688 lb: 0.731\n2.kfold0: 0.743 kfold1: 0.766 kfold2: 0.803 kfold3: 0.764 kfold4: 0.777 avg: 0.7706 lb: 0.734\n```\n\n**Finally, congratulations to all the medal winners, now you can have a good sleep.**",
      "votes": 14
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2550382": "**In this competition, we utilized the PrecTime model, which had satisfactory results. Hence, we are sharing it here with everyone.**\n\n**As the original paper did not provide open-source code, we attempted to reproduce it. However, there might be some misunderstanding of its structure, which could have prevented us from reaching its highest potential score. I believe other teams might have also used this model. Looking forward to everyone's discussion.**\n\nPaper: [PrecTime: A Deep Learning Architecture for Precise Time Series Segmentation in Industrial Manufacturing Operations](https://arxiv.org/abs/2302.10182)\nOur Code: [Here](https://github.com/Lizhecheng02/Kaggle-Detect_Sleep_States)\n\nMain Structure:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F10671485%2F854eb1755cd5e1a63104e4f4cc0ac248%2F_20231205181014.png?generation=1701828638336931&alt=media)\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F10671485%2F2f25d9b761b92ffefabad60dba87b843%2F_20231205181019.png?generation=1701828815897346&alt=media)\n\n```\nResults:\n1.kfold0: 0.743 kfold1: 0.769 kfold2: 0.793 kfold3: 0.761 kfold4: 0.772 avg: 0.7688 lb: 0.731\n2.kfold0: 0.743 kfold1: 0.766 kfold2: 0.803 kfold3: 0.764 kfold4: 0.777 avg: 0.7706 lb: 0.734\n```\n\n**Finally, congratulations to all the medal winners, now you can have a good sleep.**"
  }
}