{
  "id": 239457,
  "title": "Is string matching allowed?",
  "url": "/competitions/coleridgeinitiative-show-us-the-data/discussion/239457",
  "author_name": "",
  "post_date": "2021-05-16T12:22:00.604423600Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Most submissions seem to have achieved a public LB score of about 0.51 - 0.53, and such a score is more likely a result of string matching. Based on my reading of the discussions, non-string matching models struggle to reach the string matching benchmark.</p>\n<p>This brings a question, is string matching allowed for this competition?</p>\n<p>What approaches are other competitors taking? Is it,</p>\n<ul>\n<li>String matching alone</li>\n<li>NLP model alone</li>\n<li>String matching augmented with NLP models</li>\n</ul>",
  "messages": [
    {
      "id": "1309990",
      "postDate": "05/16/2021 12:22:00",
      "content": "<p>Most submissions seem to have achieved a public LB score of about 0.51 - 0.53, and such a score is more likely a result of string matching. Based on my reading of the discussions, non-string matching models struggle to reach the string matching benchmark.</p>\n<p>This brings a question, is string matching allowed for this competition?</p>\n<p>What approaches are other competitors taking? Is it,</p>\n<ul>\n<li>String matching alone</li>\n<li>NLP model alone</li>\n<li>String matching augmented with NLP models</li>\n</ul>",
      "rawMarkdown": "Most submissions seem to have achieved a public LB score of about 0.51 - 0.53, and such a score is more likely a result of string matching. Based on my reading of the discussions, non-string matching models struggle to reach the string matching benchmark.\n\nThis brings a question, is string matching allowed for this competition?\n\nWhat approaches are other competitors taking? Is it,\n- String matching alone\n- NLP model alone\n- String matching augmented with NLP models",
      "votes": null
    },
    {
      "id": "1310017",
      "postDate": "05/16/2021 12:41:16",
      "content": "<p>Yes allowed, but cannot generalize on unseen text/datasets.</p>",
      "rawMarkdown": "Yes allowed, but cannot generalize on unseen text/datasets.",
      "votes": null
    },
    {
      "id": "1324592",
      "postDate": "05/27/2021 04:51:43",
      "content": "<p>yes it is allowed but it doesn't seems that it will give same score on private leaderboard.</p>",
      "rawMarkdown": "yes it is allowed but it doesn't seems that it will give same score on private leaderboard.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1310017,
      "author_name": "mlconsult",
      "author_url": "",
      "post_date": "05/16/2021 12:41:16",
      "content": "<p>Yes allowed, but cannot generalize on unseen text/datasets.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1324592,
      "author_name": "ajaypawar123",
      "author_url": "",
      "post_date": "05/27/2021 04:51:43",
      "content": "<p>yes it is allowed but it doesn't seems that it will give same score on private leaderboard.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1309990": "Most submissions seem to have achieved a public LB score of about 0.51 - 0.53, and such a score is more likely a result of string matching. Based on my reading of the discussions, non-string matching models struggle to reach the string matching benchmark.\n\nThis brings a question, is string matching allowed for this competition?\n\nWhat approaches are other competitors taking? Is it,\n- String matching alone\n- NLP model alone\n- String matching augmented with NLP models",
    "1310017": "Yes allowed, but cannot generalize on unseen text/datasets.",
    "1324592": "yes it is allowed but it doesn't seems that it will give same score on private leaderboard."
  },
  "source": "meta"
}