{
  "id": 4320,
  "title": "paper: \"Detecting Causality in Complex Ecosystems\"",
  "url": "/competitions/cause-effect-pairs/discussion/4320",
  "author_name": "",
  "post_date": "2013-04-15T14:22:34.587Z",
  "votes": 2,
  "comment_count": 4,
  "views": 2625,
  "content": "<p>Gelman just <a href=\"http://andrewgelman.com/2013/04/14/detecting-predictability-in-complex-ecosystems/\">\r\nposted</a> on this paper. Possibly relevant?</p>\r\n<p><a href=\"http://www.uvm.edu/~cdanfort/csc-reading-group/sugihara-causality-science-2012.pdf\">Detecting Causality in Complex Ecosystems</a></p>",
  "messages": [
    {
      "id": "22838",
      "postDate": "04/15/2013 14:22:34",
      "content": "<p>Gelman just <a href=\"http://andrewgelman.com/2013/04/14/detecting-predictability-in-complex-ecosystems/\">\r\nposted</a> on this paper. Possibly relevant?</p>\r\n<p><a href=\"http://www.uvm.edu/~cdanfort/csc-reading-group/sugihara-causality-science-2012.pdf\">Detecting Causality in Complex Ecosystems</a></p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "22844",
      "postDate": "04/15/2013 15:18:06",
      "content": "<p>Is it for time-series as well? like Granger?</p>\r\n<p>&nbsp;</p>\r\n<p>&quot;Our alternative approach, convergent crossmapping<br>\r\n(CCM), tests for causation by measuring the<br>\r\nextent to which the historical record of Yvalues can<br>\r\nreliably estimate states of X. This happens only if X<br>\r\nis causally influencing Y. Inmoredetail, CCMlooks<br>\r\nfor the signature of X in Y’s time series by seeing<br>\r\nwhether there is a correspondence....&quot;</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "23234",
      "postDate": "04/20/2013 16:37:44",
      "content": "<p>It looks so. So, probably it is not applicable to the data sets in this competition seen that we have no info on the time structure.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "23315",
      "postDate": "04/22/2013 02:05:48",
      "content": "<p>It is my understanding that the data in this competition are not time-series.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "23913",
      "postDate": "05/05/2013 11:02:14",
      "content": "<p>great attractive paper &nbsp;:)_</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 22844,
      "author_name": "domcastro",
      "author_url": "",
      "post_date": "04/15/2013 15:18:06",
      "content": "<p>Is it for time-series as well? like Granger?</p>\r\n<p>&nbsp;</p>\r\n<p>&quot;Our alternative approach, convergent crossmapping<br>\r\n(CCM), tests for causation by measuring the<br>\r\nextent to which the historical record of Yvalues can<br>\r\nreliably estimate states of X. This happens only if X<br>\r\nis causally influencing Y. Inmoredetail, CCMlooks<br>\r\nfor the signature of X in Y’s time series by seeing<br>\r\nwhether there is a correspondence....&quot;</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 23234,
      "author_name": "larry77",
      "author_url": "",
      "post_date": "04/20/2013 16:37:44",
      "content": "<p>It looks so. So, probably it is not applicable to the data sets in this competition seen that we have no info on the time structure.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 23315,
      "author_name": "jc3626439",
      "author_url": "",
      "post_date": "04/22/2013 02:05:48",
      "content": "<p>It is my understanding that the data in this competition are not time-series.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 23913,
      "author_name": "harriken",
      "author_url": "",
      "post_date": "05/05/2013 11:02:14",
      "content": "<p>great attractive paper &nbsp;:)_</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "22838": "",
    "22844": "",
    "23234": "",
    "23315": "",
    "23913": ""
  },
  "source": "meta"
}