![\"![\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F16438831%2F09664bad1cd512ee26aa8cfea0b6d758%2FFP.JPG?generation=1741495578492483&alt=media\")
]Precision: What proportion of the positive identifications (predicted boxes) was actually correct?
\nPrecision = TP/(TP + FP)
Recall: What proportion of actual positives was identified correctly?
\nRecall = TP / (TP + FN)
]
TP: True Positives. You Predict, and it is correct.
FP: False Positives. You Predict, but it is wrong
FN: False Negative. You did not predict. and it is worng. You should have predicted
There are no true negatives(TN) in object detection
\nF1-Score: The F1 Score is the harmonic mean of precision and recall, providing a balanced assessment of a model's performance while considering both false positives and false negativs
F0.5-Score: The F0.5 Score is a variation of the F1 Score that places more emphasis on precision than recall.
F2-Score: The F2 Score is a variation of the F1 Score that places more emphasis on recall than precision.
\nIt means that competition host Focus more on the presence of actual boxes rather than the accuracy of box predictions.
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F16438831%2F52de9b51f2da84ba338cad5ddb64fb10%2Fsit0.JPG?generation=1741496681116880&alt=media\")
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F16438831%2Fea8bd3859644a30c0ada0615015c037b%2Fsit0.JPG?generation=1741496851061704&alt=media\")
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F16438831%2F858e78159de7e4e6d281915e60d01bb7%2Fsit%202.JPG?generation=1741496761707696&alt=media\")
True Positive (TP): If |y−y¯|2≤τ, the prediction is within threshold.
False Negative (FN): If |y−y¯|2>τ, the prediction is outside of threshold.
Threshold: τ=1000 Angstroms.
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F16438831%2Fe0baaff5d7aef38170df584f91b05757%2FEUCLIDE.JPG?generation=1741498287340289&alt=media\")
Precision: What proportion of the positive identifications (predicted boxes) was actually correct?
\nPrecision = TP/(TP + FP)
Recall: What proportion of actual positives was identified correctly?
\nRecall = TP / (TP + FN)
]
TP: True Positives. You Predict, and it is correct.
FP: False Positives. You Predict, but it is wrong
FN: False Negative. You did not predict. and it is worng. You should have predicted
There are no true negatives(TN) in object detection
\nF1-Score: The F1 Score is the harmonic mean of precision and recall, providing a balanced assessment of a model's performance while considering both false positives and false negativs
F0.5-Score: The F0.5 Score is a variation of the F1 Score that places more emphasis on precision than recall.
F2-Score: The F2 Score is a variation of the F1 Score that places more emphasis on recall than precision.
\nIt means that competition host Focus more on the presence of actual boxes rather than the accuracy of box predictions.
True Positive (TP): If |y−y¯|2≤τ, the prediction is within threshold.
False Negative (FN): If |y−y¯|2>τ, the prediction is outside of threshold.
Threshold: τ=1000 Angstroms.
Thanks for the explanation!
", "rawMarkdown": "Thanks for the explanation!", "votes": null } ], "comments": [ { "id": 3160544, "author_name": "", "author_url": "", "post_date": "03/26/2025 22:26:13", "content": "Thanks for the explanation!
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "3144956": "### Basics\n\n#### Precision and Recall\n**Precision**: What proportion of the positive identifications (predicted boxes) was actually correct?\n\n`Precision = TP/(TP + FP)`\n\n**Recall**: What proportion of actual positives was identified correctly?\n\n`Recall = TP / (TP + FN)`\n\n\n]\n\n**TP**: `True Positives`. You Predict, and it is correct.\n\n**FP**: `False Positives`. You Predict, but it is wrong\n\n**FN**: `False Negative`. You did not predict. and it is worng. You should have predicted\n\nThere are no true negatives(TN) in object detection\n\n### Fβ-score\n\n- **`F1-Score`**: The F1 Score is the harmonic mean of precision and recall, providing a balanced assessment of a model's performance while considering both false positives and false negativs \n\n- **`F0.5-Score`**: The F0.5 Score is a variation of the F1 Score that places more emphasis on precision than recall.\n\n\n- **`F2-Score`**: The F2 Score is a variation of the F1 Score that places more emphasis on recall than precision. \nIt means that competition host Focus more on the presence of actual boxes rather than the accuracy of box predictions.\n\n\n## Case 1: Precision = Recall\n\n\n\n## Case 2: Precision < Recall\n\n\n\n## Case 3: Precision > Recall\n\n\n\n# Euclidean distance \n\n- **True Positive (TP)**: If `|y−y¯|2≤τ`, the prediction is within threshold.\n\n- **False Negative (FN)**: If `|y−y¯|2>τ`, the prediction is outside of threshold.\n\nThreshold: `τ=1000 Angstroms.`\n\n### if |y−y¯|2>τ, Decrease TP, Increase FN\n\n", "3160544": "Thanks for the explanation!" }, "source": "meta" }