{
  "id": 433865,
  "title": "Please share your LB and CV on each of the three types of gene ontologies",
  "url": "/competitions/cafa-5-protein-function-prediction/discussion/433865",
  "author_name": "",
  "post_date": "2023-08-23T07:07:42.913137400Z",
  "votes": 2,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Hello, everyone.</p>\n<p>Let me start by sharing my LB scores on each of the three GO categories. <br>\nBPO：0.137 CCO：0.257 MFO：0.225</p>\n<p>I know I didn't do so well, so I'm very curious to see what category I can do better in.</p>",
  "messages": [
    {
      "id": "2404300",
      "postDate": "08/23/2023 07:07:42",
      "content": "<p>Hello, everyone.</p>\n<p>Let me start by sharing my LB scores on each of the three GO categories. <br>\nBPO：0.137 CCO：0.257 MFO：0.225</p>\n<p>I know I didn't do so well, so I'm very curious to see what category I can do better in.</p>",
      "rawMarkdown": "Hello, everyone.\n\nLet me start by sharing my LB scores on each of the three GO categories. \nBPO：0.137 CCO：0.257 MFO：0.225\n\nI know I didn't do so well, so I'm very curious to see what category I can do better in.",
      "votes": null
    },
    {
      "id": "2404408",
      "postDate": "08/23/2023 08:32:30",
      "content": "<p>I’m curious how did you measure LB scores for each sub-ontology? Did you try separate submissions for each of them?</p>\n<p>P.S seems hard to say that you didn’t do so well while holding 13th place at LB😅</p>",
      "rawMarkdown": "I’m curious how did you measure LB scores for each sub-ontology? Did you try separate submissions for each of them?\n\nP.S seems hard to say that you didn’t do so well while holding 13th place at LB😅",
      "votes": null
    },
    {
      "id": "2404533",
      "postDate": "08/23/2023 09:50:04",
      "content": "<p>hi,<br>\nDue to the large differences between sub-ontologies, I trained models for each sub-ontology and submitted them separately.</p>",
      "rawMarkdown": "hi,\nDue to the large differences between sub-ontologies, I trained models for each sub-ontology and submitted them separately.",
      "votes": null
    },
    {
      "id": "2404659",
      "postDate": "08/23/2023 12:25:39",
      "content": "<p>BPO: 0.14243<br>\nCCO: 0.25437<br>\nMFO: 0.22203</p>",
      "rawMarkdown": "BPO: 0.14243\nCCO: 0.25437\nMFO: 0.22203",
      "votes": null
    },
    {
      "id": "2405537",
      "postDate": "08/24/2023 00:57:31",
      "content": "<p>hi, thanks for you comment.<br>\nDo you train the model separately as well? I'm wondering if you merge the results from the top public code. Because when I try to integrate the results from Diamond or top public the LB score goes down instead.</p>",
      "rawMarkdown": "hi, thanks for you comment.\nDo you train the model separately as well? I'm wondering if you merge the results from the top public code. Because when I try to integrate the results from Diamond or top public the LB score goes down instead.",
      "votes": null
    },
    {
      "id": "2408064",
      "postDate": "08/25/2023 12:16:31",
      "content": "<p>Yes, I've trained separately for each ontology. In CC ontology, we used a modified version of one of the public code, that showed best results compared to the standard code</p>",
      "rawMarkdown": "Yes, I've trained separately for each ontology. In CC ontology, we used a modified version of one of the public code, that showed best results compared to the standard code",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2404408,
      "author_name": "aucrock",
      "author_url": "",
      "post_date": "08/23/2023 08:32:30",
      "content": "<p>I’m curious how did you measure LB scores for each sub-ontology? Did you try separate submissions for each of them?</p>\n<p>P.S seems hard to say that you didn’t do so well while holding 13th place at LB😅</p>",
      "votes": null,
      "replies": [
        {
          "id": 2404533,
          "author_name": "impimp",
          "author_url": "",
          "post_date": "08/23/2023 09:50:04",
          "content": "<p>hi,<br>\nDue to the large differences between sub-ontologies, I trained models for each sub-ontology and submitted them separately.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2404659,
      "author_name": "gabrielbianchin",
      "author_url": "",
      "post_date": "08/23/2023 12:25:39",
      "content": "<p>BPO: 0.14243<br>\nCCO: 0.25437<br>\nMFO: 0.22203</p>",
      "votes": null,
      "replies": [
        {
          "id": 2405537,
          "author_name": "impimp",
          "author_url": "",
          "post_date": "08/24/2023 00:57:31",
          "content": "<p>hi, thanks for you comment.<br>\nDo you train the model separately as well? I'm wondering if you merge the results from the top public code. Because when I try to integrate the results from Diamond or top public the LB score goes down instead.</p>",
          "votes": null,
          "replies": [
            {
              "id": 2408064,
              "author_name": "gabrielbianchin",
              "author_url": "",
              "post_date": "08/25/2023 12:16:31",
              "content": "<p>Yes, I've trained separately for each ontology. In CC ontology, we used a modified version of one of the public code, that showed best results compared to the standard code</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2404300": "Hello, everyone.\n\nLet me start by sharing my LB scores on each of the three GO categories. \nBPO：0.137 CCO：0.257 MFO：0.225\n\nI know I didn't do so well, so I'm very curious to see what category I can do better in.",
    "2404408": "I’m curious how did you measure LB scores for each sub-ontology? Did you try separate submissions for each of them?\n\nP.S seems hard to say that you didn’t do so well while holding 13th place at LB😅",
    "2404533": "hi,\nDue to the large differences between sub-ontologies, I trained models for each sub-ontology and submitted them separately.",
    "2404659": "BPO: 0.14243\nCCO: 0.25437\nMFO: 0.22203",
    "2405537": "hi, thanks for you comment.\nDo you train the model separately as well? I'm wondering if you merge the results from the top public code. Because when I try to integrate the results from Diamond or top public the LB score goes down instead.",
    "2408064": "Yes, I've trained separately for each ontology. In CC ontology, we used a modified version of one of the public code, that showed best results compared to the standard code"
  },
  "source": "meta"
}