{
  "id": 7537,
  "title": "AUC in training?",
  "url": "/competitions/connectomics/discussion/7537",
  "author_name": "",
  "post_date": "2014-03-28T13:46:00.907Z",
  "votes": null,
  "comment_count": 1,
  "views": 781,
  "content": "<p>I am having some issues running the sample code<br><br>Could anyone tell me what AUC is obtained with crossCorrelation, GTE, etc on normal-1&nbsp;network (file: fluorescence_normal1-1.txt) to check my results?<br><br>I am currently getting AUC = 0.57 with crossCorrelation and discretization, which is not very promising...</p>",
  "messages": [
    {
      "id": "41126",
      "postDate": "03/28/2014 13:46:00",
      "content": "<p>I am having some issues running the sample code<br><br>Could anyone tell me what AUC is obtained with crossCorrelation, GTE, etc on normal-1&nbsp;network (file: fluorescence_normal1-1.txt) to check my results?<br><br>I am currently getting AUC = 0.57 with crossCorrelation and discretization, which is not very promising...</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "41227",
      "postDate": "03/31/2014 13:45:39",
      "content": "<p>Hmm, I think cross-correlation with 3 bin discretization in normal-1 gives around 0.67. For testing code I would recomend to work on the small data set, since it only has 100 neurons.&nbsp;</p>\n<p>GTE on the iNet1_Size100_CC06inh dataset should be around 0.92, and cross-correlation around&nbsp;0.88.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 41227,
      "author_name": "orlandi",
      "author_url": "",
      "post_date": "03/31/2014 13:45:39",
      "content": "<p>Hmm, I think cross-correlation with 3 bin discretization in normal-1 gives around 0.67. For testing code I would recomend to work on the small data set, since it only has 100 neurons.&nbsp;</p>\n<p>GTE on the iNet1_Size100_CC06inh dataset should be around 0.92, and cross-correlation around&nbsp;0.88.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "41126": "",
    "41227": ""
  },
  "source": "meta"
}