{
  "id": 403178,
  "title": "667th place, but still happy :)",
  "url": "/competitions/icecube-neutrinos-in-deep-ice/discussion/403178",
  "author_name": "",
  "post_date": "2023-04-21T16:21:33.874582400Z",
  "votes": 14,
  "comment_count": 3,
  "views": 0,
  "content": "<p>I would like to give great thanks to the organizers of this competition. The chance to model simulated hits at the same time to try and learn a whole lot about modelling software does not happen often.</p>\n<p>For the users of <a href=\"https://crumbsoftware.com:3028/meta.htm?object=icecube\" target=\"_blank\">https://crumbsoftware.com:3028/meta.htm?object=icecube</a>, if you have any ideas or suggestions on the software, they would be appreciated.</p>\n<p>Regarding the contest. I concentrated on single string hits (usually only about 3 or 4 sensors go off, all on the same string). Unfortunately, any model seemed to give me a worse score then the starting point of a copy of the original EDA+PCA notebook. Did anyone have any luck modelling single string hits? If so, how?</p>",
  "messages": [
    {
      "id": "2229719",
      "postDate": "04/21/2023 16:21:33",
      "content": "<p>I would like to give great thanks to the organizers of this competition. The chance to model simulated hits at the same time to try and learn a whole lot about modelling software does not happen often.</p>\n<p>For the users of <a href=\"https://crumbsoftware.com:3028/meta.htm?object=icecube\" target=\"_blank\">https://crumbsoftware.com:3028/meta.htm?object=icecube</a>, if you have any ideas or suggestions on the software, they would be appreciated.</p>\n<p>Regarding the contest. I concentrated on single string hits (usually only about 3 or 4 sensors go off, all on the same string). Unfortunately, any model seemed to give me a worse score then the starting point of a copy of the original EDA+PCA notebook. Did anyone have any luck modelling single string hits? If so, how?</p>",
      "rawMarkdown": "I would like to give great thanks to the organizers of this competition. The chance to model simulated hits at the same time to try and learn a whole lot about modelling software does not happen often.\n\nFor the users of [https://crumbsoftware.com:3028/meta.htm?object=icecube](https://crumbsoftware.com:3028/meta.htm?object=icecube), if you have any ideas or suggestions on the software, they would be appreciated.\n\nRegarding the contest. I concentrated on single string hits (usually only about 3 or 4 sensors go off, all on the same string). Unfortunately, any model seemed to give me a worse score then the starting point of a copy of the original EDA+PCA notebook. Did anyone have any luck modelling single string hits? If so, how?",
      "votes": null
    },
    {
      "id": "2229900",
      "postDate": "04/21/2023 19:57:59",
      "content": "<p>Thank you for your tool. I found it incredibly useful and used it a lot. Single string hits were a big challenge for everybody</p>",
      "rawMarkdown": "Thank you for your tool. I found it incredibly useful and used it a lot. Single string hits were a big challenge for everybody",
      "votes": null
    },
    {
      "id": "2230409",
      "postDate": "04/22/2023 10:47:45",
      "content": "<p>I like your post! Enjoy journey and learning process -&gt; bronze, silver, gold … will come.</p>",
      "rawMarkdown": "I like your post! Enjoy journey and learning process -> bronze, silver, gold ... will come.",
      "votes": null
    },
    {
      "id": "2232924",
      "postDate": "04/24/2023 17:41:27",
      "content": "<p>Came here to say thanks, <a href=\"https://www.kaggle.com/edguy99\" target=\"_blank\">@edguy99</a>, your tool was a great help in exploring the dataset. I probably won't be the first to suggest, but anyway, <a href=\"https://www.kaggle.com/sohier\" target=\"_blank\">@sohier</a> would it be nice to have a regular prize or kaggle points for this kind of work?</p>\n<p>As for single-string events, well, I believe you picked hardest part of the problem. Our metric is combination of errors in zenith and azimuth angles, and azimuth is almost impossible to predict for single-string events, as they are symmetric under rotation around z axis.</p>",
      "rawMarkdown": "Came here to say thanks, @edguy99, your tool was a great help in exploring the dataset. I probably won't be the first to suggest, but anyway, @sohier would it be nice to have a regular prize or kaggle points for this kind of work?\n\nAs for single-string events, well, I believe you picked hardest part of the problem. Our metric is combination of errors in zenith and azimuth angles, and azimuth is almost impossible to predict for single-string events, as they are symmetric under rotation around z axis.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2229900,
      "author_name": "anjum48",
      "author_url": "",
      "post_date": "04/21/2023 19:57:59",
      "content": "<p>Thank you for your tool. I found it incredibly useful and used it a lot. Single string hits were a big challenge for everybody</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2230409,
      "author_name": "remekkinas",
      "author_url": "",
      "post_date": "04/22/2023 10:47:45",
      "content": "<p>I like your post! Enjoy journey and learning process -&gt; bronze, silver, gold … will come.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2232924,
      "author_name": "alexz0",
      "author_url": "",
      "post_date": "04/24/2023 17:41:27",
      "content": "<p>Came here to say thanks, <a href=\"https://www.kaggle.com/edguy99\" target=\"_blank\">@edguy99</a>, your tool was a great help in exploring the dataset. I probably won't be the first to suggest, but anyway, <a href=\"https://www.kaggle.com/sohier\" target=\"_blank\">@sohier</a> would it be nice to have a regular prize or kaggle points for this kind of work?</p>\n<p>As for single-string events, well, I believe you picked hardest part of the problem. Our metric is combination of errors in zenith and azimuth angles, and azimuth is almost impossible to predict for single-string events, as they are symmetric under rotation around z axis.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2229719": "I would like to give great thanks to the organizers of this competition. The chance to model simulated hits at the same time to try and learn a whole lot about modelling software does not happen often.\n\nFor the users of [https://crumbsoftware.com:3028/meta.htm?object=icecube](https://crumbsoftware.com:3028/meta.htm?object=icecube), if you have any ideas or suggestions on the software, they would be appreciated.\n\nRegarding the contest. I concentrated on single string hits (usually only about 3 or 4 sensors go off, all on the same string). Unfortunately, any model seemed to give me a worse score then the starting point of a copy of the original EDA+PCA notebook. Did anyone have any luck modelling single string hits? If so, how?",
    "2229900": "Thank you for your tool. I found it incredibly useful and used it a lot. Single string hits were a big challenge for everybody",
    "2230409": "I like your post! Enjoy journey and learning process -> bronze, silver, gold ... will come.",
    "2232924": "Came here to say thanks, @edguy99, your tool was a great help in exploring the dataset. I probably won't be the first to suggest, but anyway, @sohier would it be nice to have a regular prize or kaggle points for this kind of work?\n\nAs for single-string events, well, I believe you picked hardest part of the problem. Our metric is combination of errors in zenith and azimuth angles, and azimuth is almost impossible to predict for single-string events, as they are symmetric under rotation around z axis."
  },
  "source": "meta"
}