{
  "id": 376004,
  "title": "How to estimate the performance of baseline detector provided in train and test datasets?",
  "url": "/competitions/nfl-player-contact-detection/discussion/376004",
  "author_name": "",
  "post_date": "2023-01-04T11:20:41.559130300Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Different from last year's competition, this year's boxes are assigned by the last year's winner model.</p>",
  "messages": [
    {
      "id": "2085750",
      "postDate": "01/04/2023 11:20:41",
      "content": "<p>Different from last year's competition, this year's boxes are assigned by the last year's winner model.</p>",
      "rawMarkdown": "Different from last year's competition, this year's boxes are assigned by the last year's winner model.",
      "votes": null
    },
    {
      "id": "2087202",
      "postDate": "01/05/2023 12:14:35",
      "content": "<p>To estimate the performance of a baseline detector provided in the train and test datasets, you can use a metric such as the Matthews Correlation Coefficient (MCC). The MCC is a measure of the quality of binary classification, and is defined as the correlation between the predicted and observed binary classifications. It takes into account true and false positives and negatives, and ranges from -1 (perfect disagreement between prediction and observation) to +1 (perfect agreement).</p>\n<p>To calculate the MCC for the baseline detector, you will need to compare the predicted and actual contact events for each contact_id in the test set. You can then use the following formula to calculate the MCC:</p>\n<p>MCC = (TPTN - FPFN) / sqrt((TP+FP)(TP+FN)(TN+FP)*(TN+FN))</p>\n<p>Where TP is the number of true positives (correctly predicted contact events), TN is the number of true negatives (correctly predicted non-contact events), FP is the number of false positives (incorrectly predicted contact events), and FN is the number of false negatives (incorrectly predicted non-contact events).</p>\n<p>Once you have calculated the MCC, you can use it to assess the performance of the baseline detector. A MCC of 0 indicates that the detector is no better than random chance, while a MCC greater than 0 indicates that the detector is making accurate predictions. The higher the MCC, the better the performance of the detector.</p>",
      "rawMarkdown": "To estimate the performance of a baseline detector provided in the train and test datasets, you can use a metric such as the Matthews Correlation Coefficient (MCC). The MCC is a measure of the quality of binary classification, and is defined as the correlation between the predicted and observed binary classifications. It takes into account true and false positives and negatives, and ranges from -1 (perfect disagreement between prediction and observation) to +1 (perfect agreement).\n\nTo calculate the MCC for the baseline detector, you will need to compare the predicted and actual contact events for each contact_id in the test set. You can then use the following formula to calculate the MCC:\n\nMCC = (TPTN - FPFN) / sqrt((TP+FP)(TP+FN)(TN+FP)*(TN+FN))\n\nWhere TP is the number of true positives (correctly predicted contact events), TN is the number of true negatives (correctly predicted non-contact events), FP is the number of false positives (incorrectly predicted contact events), and FN is the number of false negatives (incorrectly predicted non-contact events).\n\nOnce you have calculated the MCC, you can use it to assess the performance of the baseline detector. A MCC of 0 indicates that the detector is no better than random chance, while a MCC greater than 0 indicates that the detector is making accurate predictions. The higher the MCC, the better the performance of the detector.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2087202,
      "author_name": "aviralmishra1998",
      "author_url": "",
      "post_date": "01/05/2023 12:14:35",
      "content": "<p>To estimate the performance of a baseline detector provided in the train and test datasets, you can use a metric such as the Matthews Correlation Coefficient (MCC). The MCC is a measure of the quality of binary classification, and is defined as the correlation between the predicted and observed binary classifications. It takes into account true and false positives and negatives, and ranges from -1 (perfect disagreement between prediction and observation) to +1 (perfect agreement).</p>\n<p>To calculate the MCC for the baseline detector, you will need to compare the predicted and actual contact events for each contact_id in the test set. You can then use the following formula to calculate the MCC:</p>\n<p>MCC = (TPTN - FPFN) / sqrt((TP+FP)(TP+FN)(TN+FP)*(TN+FN))</p>\n<p>Where TP is the number of true positives (correctly predicted contact events), TN is the number of true negatives (correctly predicted non-contact events), FP is the number of false positives (incorrectly predicted contact events), and FN is the number of false negatives (incorrectly predicted non-contact events).</p>\n<p>Once you have calculated the MCC, you can use it to assess the performance of the baseline detector. A MCC of 0 indicates that the detector is no better than random chance, while a MCC greater than 0 indicates that the detector is making accurate predictions. The higher the MCC, the better the performance of the detector.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2085750": "Different from last year's competition, this year's boxes are assigned by the last year's winner model.",
    "2087202": "To estimate the performance of a baseline detector provided in the train and test datasets, you can use a metric such as the Matthews Correlation Coefficient (MCC). The MCC is a measure of the quality of binary classification, and is defined as the correlation between the predicted and observed binary classifications. It takes into account true and false positives and negatives, and ranges from -1 (perfect disagreement between prediction and observation) to +1 (perfect agreement).\n\nTo calculate the MCC for the baseline detector, you will need to compare the predicted and actual contact events for each contact_id in the test set. You can then use the following formula to calculate the MCC:\n\nMCC = (TPTN - FPFN) / sqrt((TP+FP)(TP+FN)(TN+FP)*(TN+FN))\n\nWhere TP is the number of true positives (correctly predicted contact events), TN is the number of true negatives (correctly predicted non-contact events), FP is the number of false positives (incorrectly predicted contact events), and FN is the number of false negatives (incorrectly predicted non-contact events).\n\nOnce you have calculated the MCC, you can use it to assess the performance of the baseline detector. A MCC of 0 indicates that the detector is no better than random chance, while a MCC greater than 0 indicates that the detector is making accurate predictions. The higher the MCC, the better the performance of the detector."
  },
  "source": "meta"
}