{
  "id": 77051,
  "title": "How to calculate Fractal dimension",
  "url": "/competitions/vsb-power-line-fault-detection/discussion/77051",
  "author_name": "",
  "post_date": "2019-01-09T03:36:01.261104700Z",
  "votes": 2,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hi all, </p>\n\n<p>In the thesis shared over here : \n<a href=\"https://www.kaggle.com/c/vsb-power-line-fault-detection/discussion/75771\">https://www.kaggle.com/c/vsb-power-line-fault-detection/discussion/75771</a>, it was mentioned that fractal dimension was one of feature, can anyone explain how to calculate this feature.</p>\n\n<p>Thanks!</p>",
  "messages": [
    {
      "id": "452689",
      "postDate": "01/09/2019 03:36:01",
      "content": "<p>Hi all, </p>\n\n<p>In the thesis shared over here : \n<a href=\"https://www.kaggle.com/c/vsb-power-line-fault-detection/discussion/75771\">https://www.kaggle.com/c/vsb-power-line-fault-detection/discussion/75771</a>, it was mentioned that fractal dimension was one of feature, can anyone explain how to calculate this feature.</p>\n\n<p>Thanks!</p>",
      "rawMarkdown": "Hi all, \n\nIn the thesis shared over here : \nhttps://www.kaggle.com/c/vsb-power-line-fault-detection/discussion/75771, it was mentioned that fractal dimension was one of feature, can anyone explain how to calculate this feature.\n\nThanks!",
      "votes": null
    },
    {
      "id": "453219",
      "postDate": "01/09/2019 21:38:56",
      "content": "<p>I'm not going to attempt to explain it but there is an explanation of it in the paper that was cited. You can find it here <a href=\"https://www.hindawi.com/journals/jece/2015/174538/\">https://www.hindawi.com/journals/jece/2015/174538/</a></p>",
      "rawMarkdown": "I'm not going to attempt to explain it but there is an explanation of it in the paper that was cited. You can find it here https://www.hindawi.com/journals/jece/2015/174538/",
      "votes": null
    },
    {
      "id": "453234",
      "postDate": "01/09/2019 22:38:23",
      "content": "<p>Thanks Jack, I will look into the mentioned paper.</p>",
      "rawMarkdown": "Thanks Jack, I will look into the mentioned paper.",
      "votes": null
    },
    {
      "id": "453805",
      "postDate": "01/10/2019 20:13:53",
      "content": "<p>Read section 4.3: <a href=\"https://www.labri.fr/perso/nrougier/from-python-to-numpy/#temporal-vectorization\">https://www.labri.fr/perso/nrougier/from-python-to-numpy/#temporal-vectorization</a></p>\n\n<p>Then review python code here: <a href=\"https://www.labri.fr/perso/nrougier/from-python-to-numpy/code/fractal_dimension.py\">https://www.labri.fr/perso/nrougier/from-python-to-numpy/code/fractal_dimension.py</a></p>\n\n<p>That stated, given this feature's MI value (0.2820, which is less information acquired than just feeding the mean value of the signal as an input!), I'd advise finding other ways to introduce data to your models.</p>",
      "rawMarkdown": "Read section 4.3: https://www.labri.fr/perso/nrougier/from-python-to-numpy/#temporal-vectorization\n\nThen review python code here: https://www.labri.fr/perso/nrougier/from-python-to-numpy/code/fractal_dimension.py\n\nThat stated, given this feature's MI value (0.2820, which is less information acquired than just feeding the mean value of the signal as an input!), I'd advise finding other ways to introduce data to your models.",
      "votes": null
    },
    {
      "id": "454011",
      "postDate": "01/11/2019 02:51:13",
      "content": "<p>Thank you <a href=\"/authman\">@authman</a>.</p>",
      "rawMarkdown": "Thank you @authman.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 453219,
      "author_name": "jackvial",
      "author_url": "",
      "post_date": "01/09/2019 21:38:56",
      "content": "<p>I'm not going to attempt to explain it but there is an explanation of it in the paper that was cited. You can find it here <a href=\"https://www.hindawi.com/journals/jece/2015/174538/\">https://www.hindawi.com/journals/jece/2015/174538/</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 453234,
          "author_name": "harshit92",
          "author_url": "",
          "post_date": "01/09/2019 22:38:23",
          "content": "<p>Thanks Jack, I will look into the mentioned paper.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 453805,
      "author_name": "authman",
      "author_url": "",
      "post_date": "01/10/2019 20:13:53",
      "content": "<p>Read section 4.3: <a href=\"https://www.labri.fr/perso/nrougier/from-python-to-numpy/#temporal-vectorization\">https://www.labri.fr/perso/nrougier/from-python-to-numpy/#temporal-vectorization</a></p>\n\n<p>Then review python code here: <a href=\"https://www.labri.fr/perso/nrougier/from-python-to-numpy/code/fractal_dimension.py\">https://www.labri.fr/perso/nrougier/from-python-to-numpy/code/fractal_dimension.py</a></p>\n\n<p>That stated, given this feature's MI value (0.2820, which is less information acquired than just feeding the mean value of the signal as an input!), I'd advise finding other ways to introduce data to your models.</p>",
      "votes": null,
      "replies": [
        {
          "id": 454011,
          "author_name": "harshit92",
          "author_url": "",
          "post_date": "01/11/2019 02:51:13",
          "content": "<p>Thank you <a href=\"/authman\">@authman</a>.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "452689": "Hi all, \n\nIn the thesis shared over here : \nhttps://www.kaggle.com/c/vsb-power-line-fault-detection/discussion/75771, it was mentioned that fractal dimension was one of feature, can anyone explain how to calculate this feature.\n\nThanks!",
    "453219": "I'm not going to attempt to explain it but there is an explanation of it in the paper that was cited. You can find it here https://www.hindawi.com/journals/jece/2015/174538/",
    "453234": "Thanks Jack, I will look into the mentioned paper.",
    "453805": "Read section 4.3: https://www.labri.fr/perso/nrougier/from-python-to-numpy/#temporal-vectorization\n\nThen review python code here: https://www.labri.fr/perso/nrougier/from-python-to-numpy/code/fractal_dimension.py\n\nThat stated, given this feature's MI value (0.2820, which is less information acquired than just feeding the mean value of the signal as an input!), I'd advise finding other ways to introduce data to your models.",
    "454011": "Thank you @authman."
  },
  "source": "meta"
}