Enhancing XGBoost model accuracy for nutritional profile prediction

[langbart]

Background

We have implemented an XGBoost model to predict the nutritional profiles of fishing catches based on various fishing strategies. While the model has shown promising results, there is a noticeable disparity in accuracy between models trained on data from the mainland and those from Atauro. For more detailed information and visual representation, refer to the repository book and the plot below:

Objective

The primary goal is to improve the performance of mainland XGBoost models (if applicable), and to investigate the underlying causes of the observed accuracy discrepancy between mainland and Atauro models.