In this notebook we take a look at a simple binary classification task. We want to determine whether a Pokémon is legendary or not based on its features such as effectiveness against other Pokémon types, its weight, and more.
We achieve an accuracy of 98% and an F1-score for non-legendary Pokémon of 0.99 and for legendary classification of 0.94!
The difficulty with this dataset is that it is heavily imbalanced: ~90% non-legendary Pokémon and ~10% legendary Pokémon. By implementing the regularization techniques data standardization, stratification on train-test split, batch normalization, and dropout we can combat this imbalance and achieve the results mentioned above!