### Evaluation Metric

Submissions are evaluated using **Mean (macro-averaged) Average Precision (mAP)** over all nine classes. The final score is the arithmetic mean of the Average Precision (AP) calculated for each of the 9 columns (Clutter + 8 Bird Species).

The final score is calculated as:

$$
\text{Score} = \frac{1}{9}\sum_{i=1}^{9} AP_{i}
$$

Where  \\(AP_{i}\\) is the Average Precision for class \\(i\\). The nine classes are:
1.  **Clutter**
2.  **Cormorants**
3.  **Pigeons**
4.  **Ducks**
5.  **Geese**
6.  **Gulls**
7.  **Birds of Prey**
8.  **Waders**
9.  **Songbirds**

*Note: For Average Precision we use the [SK-learn implementation](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.average_precision_score.html).*

A notebook with the implementation of the metric can be found in the following [link](https://www.kaggle.com/code/willemdieleman/ai-cup-metric).

### Submission Format
Submissions in this competition must be done by uploading a single CSV or Parquet file. Each row corresponds to a unique radar `track_id`. The **required columns** are: 
* `track_id`: The unique identifier for the track.
* **9 Class Columns**: A predicted probability (float between 0.0 and 1.0) for each of the 9 classes listed above.

For an exact template, refer to `sample_submission.csv` in the Data section.