Use supervised machine learning to predict the sex of individual seabirds based on body measurements.
The Laysan Albatross (Phoebastria immutabilis) is a large seabird found throughout the North Pacific. Unlike many bird species, male and female Laysan Albatrosses look almost identical. Because of this, researchers studying albatross populations, breeding behavior, and conservation status often struggle to determine an individual bird's sex without invasive sampling (e.g., blood tests) or direct behavioral observation.
However, subtle morphological differences — such as bill length, bill depth, wing chord, and body mass — do exist between males and females on average. This competition challenges you to build a model that predicts the sex of a Laysan Albatross from a set of body measurements, offering a fast, low-cost, and non-invasive alternative to lab-based sexing methods.
The data for this competition was collected from the Guadalupe Island seabird colony, off the coast of Baja California, Mexico, by Grupo de Ecología y Conservación de Islas (GECI), a Mexican non-profit organization dedicated to the conservation and restoration of island ecosystems. GECI's long-term seabird monitoring programs provide critical data for protecting nesting colonies across Mexico's Pacific islands.
Accurate, accessible sex-prediction tools like this can support field researchers and conservationists working to monitor and protect this near-threatened species.
This competition is a classic Prediction Competition. You will develop your model using a provided training dataset, generate predictions on a test set, and submit your results for evaluation on our leaderboard.
Your goal is to predict whether a given Laysan Albatross is male (M) or female (H) based on a set of morphometric measurements. For each bird in the test set, you must predict its sex.
Submission will be measured by the F1 Score of your predictions on a hidden test set. The F1 Score is an excellent choice for this task as it balances precision and recall, providing a robust measure of performance.
Submissions are evaluated using the F1 Score, which is the harmonic mean of precision and recall. This metric is defined as:
For this binary classification task, the F1 Score is calculated. The score ranges from 0 to 1, with 1 representing perfect precision and recall.
Public Leaderboard: Based on a subset of the test data, visible during the competition.
Private Leaderboard: Based on the remainder of the test data, used to determine the final rankings after the competition closes.
Avoid "chasing" the public leaderboard, as overfitting to that subset may not reflect performance on the private data.