Figure 7

The impact of an imbalance of the training dataset. Evaluation results on training datasets of different imbalance levels by changing the ratio between positive and negative samples. The impact of dataset imbalance was evaluated based on the false positive (FP) and false negative (FN) rates.