Which of the following risks is data augmentation MOST effective at helping to reduce?
Data augmentation increases the size and diversity of a training dataset by generating modified copies of existing samples (e.g., rotation, flipping, cropping, noise addition). This exposes the model to a wider variety of examples during training, which reduces the model's tendency to memorize the training set and improves its ability to generalize to new, unseen data — directly mitigating overfitting. Feature scaling and normalization are unrelated preprocessing steps that adjust the numeric range of input features and are not addressed by augmenting the dataset.
Community Discussion