{
  "id": 402616,
  "title": "How much score did you get with train data alone?",
  "url": "/competitions/tlvmc-parkinsons-freezing-gait-prediction/discussion/402616",
  "author_name": "",
  "post_date": "2023-04-19T03:02:19.854168800Z",
  "votes": 13,
  "comment_count": 10,
  "views": 0,
  "content": "<p>I got the following scores with train data (<strong>train/tdcsfog</strong> and <strong>train/defog</strong>) alone.</p>\n<ul>\n<li>CV:0.403</li>\n<li>LB:0.359</li>\n</ul>\n<p>I have not used <strong>train/notype</strong> and <strong>unlabel data</strong> yet.<br>\nHow much score did you get with train data (train/tdcsfog and train/defog) <strong>alone</strong>?<br>\n<em>Note: Using the ensemble, I improved best score (LB: 0.359).</em></p>",
  "messages": [
    {
      "id": "2226536",
      "postDate": "04/19/2023 03:02:19",
      "content": "<p>I got the following scores with train data (<strong>train/tdcsfog</strong> and <strong>train/defog</strong>) alone.</p>\n<ul>\n<li>CV:0.403</li>\n<li>LB:0.359</li>\n</ul>\n<p>I have not used <strong>train/notype</strong> and <strong>unlabel data</strong> yet.<br>\nHow much score did you get with train data (train/tdcsfog and train/defog) <strong>alone</strong>?<br>\n<em>Note: Using the ensemble, I improved best score (LB: 0.359).</em></p>",
      "rawMarkdown": "I got the following scores with train data (**train/tdcsfog** and **train/defog**) alone.\n+ CV:0.403\n+ LB:~~0.346~~0.359\n\nI have not used **train/notype** and **unlabel data** yet.\nHow much score did you get with train data (train/tdcsfog and train/defog) **alone**?\n*Note: Using the ensemble, I improved best score (LB: 0.359).*",
      "votes": null
    },
    {
      "id": "2226657",
      "postDate": "04/19/2023 05:39:15",
      "content": "<p>I got <br>\ncv:0.440<br>\nLB:0.431<br>\nusing CNN model</p>",
      "rawMarkdown": "I got \ncv:0.440\nLB:0.431\nusing CNN model",
      "votes": null
    },
    {
      "id": "2226827",
      "postDate": "04/19/2023 09:07:03",
      "content": "<p>Would you mind to tell which tech you use? ml or dl👀</p>",
      "rawMarkdown": "Would you mind to tell which tech you use? ml or dl👀",
      "votes": null
    },
    {
      "id": "2227275",
      "postDate": "04/19/2023 16:04:59",
      "content": "<p>cv: 0.32<br>\nLB: 0.304 </p>",
      "rawMarkdown": "cv: 0.32\nLB: 0.304",
      "votes": null
    },
    {
      "id": "2227819",
      "postDate": "04/20/2023 03:39:27",
      "content": "<p>I used LightGBM, NN and CNN.<br>\nAbove score is these model's ensemble result.</p>",
      "rawMarkdown": "I used LightGBM, NN and CNN.\nAbove score is these model's ensemble result.",
      "votes": null
    },
    {
      "id": "2227820",
      "postDate": "04/20/2023 03:40:01",
      "content": "<p>I updated my result.</p>\n<blockquote>\n  <p>Note: Using the ensemble, I improved best score (LB: 0.359).</p>\n</blockquote>",
      "rawMarkdown": "I updated my result.\n> Note: Using the ensemble, I improved best score (LB: 0.359).",
      "votes": null
    },
    {
      "id": "2227943",
      "postDate": "04/20/2023 06:17:32",
      "content": "<p>thinks for shaing👀</p>",
      "rawMarkdown": "thinks for shaing👀",
      "votes": null
    },
    {
      "id": "2228737",
      "postDate": "04/20/2023 18:54:24",
      "content": "<p>would you mind to tell the data feature processing ? tsflex or rolling window ?  </p>",
      "rawMarkdown": "would you mind to tell the data feature processing ? tsflex or rolling window ?",
      "votes": null
    },
    {
      "id": "2229525",
      "postDate": "04/21/2023 12:54:39",
      "content": "<p>Anyone using DL for serieses? (LSTMs etc?)</p>\n<p>The Devastator.</p>",
      "rawMarkdown": "Anyone using DL for serieses? (LSTMs etc?)\n\n\nThe Devastator.",
      "votes": null
    },
    {
      "id": "2229683",
      "postDate": "04/21/2023 15:37:43",
      "content": "<p>Hi,  can i ask that<br>\ndo you mix all tdcsfog and defog data to train model, or you separate modeling ?<br>\nthx👍</p>",
      "rawMarkdown": "Hi,  can i ask that\ndo you mix all tdcsfog and defog data to train model, or you separate modeling ?\nthx👍",
      "votes": null
    },
    {
      "id": "2258207",
      "postDate": "05/14/2023 01:52:30",
      "content": "<p>Hi SHINMURA0, how did you calculate mAP? because sklearn function only support binary calssification. Did you train 3 binary classification models and then used sklearn to calculate mAP?</p>",
      "rawMarkdown": "Hi SHINMURA0, how did you calculate mAP? because sklearn function only support binary calssification. Did you train 3 binary classification models and then used sklearn to calculate mAP?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2226657,
      "author_name": "hiroakifukuse",
      "author_url": "",
      "post_date": "04/19/2023 05:39:15",
      "content": "<p>I got <br>\ncv:0.440<br>\nLB:0.431<br>\nusing CNN model</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2226827,
      "author_name": "roger92",
      "author_url": "",
      "post_date": "04/19/2023 09:07:03",
      "content": "<p>Would you mind to tell which tech you use? ml or dl👀</p>",
      "votes": null,
      "replies": [
        {
          "id": 2227819,
          "author_name": "shinmurashinmura",
          "author_url": "",
          "post_date": "04/20/2023 03:39:27",
          "content": "<p>I used LightGBM, NN and CNN.<br>\nAbove score is these model's ensemble result.</p>",
          "votes": null,
          "replies": [
            {
              "id": 2227943,
              "author_name": "roger92",
              "author_url": "",
              "post_date": "04/20/2023 06:17:32",
              "content": "<p>thinks for shaing👀</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    },
    {
      "id": 2227275,
      "author_name": "avivlevi815",
      "author_url": "",
      "post_date": "04/19/2023 16:04:59",
      "content": "<p>cv: 0.32<br>\nLB: 0.304 </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2227820,
      "author_name": "shinmurashinmura",
      "author_url": "",
      "post_date": "04/20/2023 03:40:01",
      "content": "<p>I updated my result.</p>\n<blockquote>\n  <p>Note: Using the ensemble, I improved best score (LB: 0.359).</p>\n</blockquote>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2228737,
      "author_name": "fakeroggy",
      "author_url": "",
      "post_date": "04/20/2023 18:54:24",
      "content": "<p>would you mind to tell the data feature processing ? tsflex or rolling window ?  </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2229525,
      "author_name": "thedevastator",
      "author_url": "",
      "post_date": "04/21/2023 12:54:39",
      "content": "<p>Anyone using DL for serieses? (LSTMs etc?)</p>\n<p>The Devastator.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2229683,
      "author_name": "algeryang",
      "author_url": "",
      "post_date": "04/21/2023 15:37:43",
      "content": "<p>Hi,  can i ask that<br>\ndo you mix all tdcsfog and defog data to train model, or you separate modeling ?<br>\nthx👍</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2258207,
      "author_name": "aidvpr",
      "author_url": "",
      "post_date": "05/14/2023 01:52:30",
      "content": "<p>Hi SHINMURA0, how did you calculate mAP? because sklearn function only support binary calssification. Did you train 3 binary classification models and then used sklearn to calculate mAP?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2226536": "I got the following scores with train data (**train/tdcsfog** and **train/defog**) alone.\n+ CV:0.403\n+ LB:~~0.346~~0.359\n\nI have not used **train/notype** and **unlabel data** yet.\nHow much score did you get with train data (train/tdcsfog and train/defog) **alone**?\n*Note: Using the ensemble, I improved best score (LB: 0.359).*",
    "2226657": "I got \ncv:0.440\nLB:0.431\nusing CNN model",
    "2226827": "Would you mind to tell which tech you use? ml or dl👀",
    "2227275": "cv: 0.32\nLB: 0.304",
    "2227819": "I used LightGBM, NN and CNN.\nAbove score is these model's ensemble result.",
    "2227820": "I updated my result.\n> Note: Using the ensemble, I improved best score (LB: 0.359).",
    "2227943": "thinks for shaing👀",
    "2228737": "would you mind to tell the data feature processing ? tsflex or rolling window ?",
    "2229525": "Anyone using DL for serieses? (LSTMs etc?)\n\n\nThe Devastator.",
    "2229683": "Hi,  can i ask that\ndo you mix all tdcsfog and defog data to train model, or you separate modeling ?\nthx👍",
    "2258207": "Hi SHINMURA0, how did you calculate mAP? because sklearn function only support binary calssification. Did you train 3 binary classification models and then used sklearn to calculate mAP?"
  },
  "source": "meta"
}