HormoNet: a deep learning approach for hormone-drug interaction prediction

BMC Bioinformatics

Table 2 Our model’s performance before and after applying SMOTE technique

Model		Accuracy	F1-Score	Precision	Recall
Train	Before	0.7773	0.3570	0.7009	0.3682
Train	After	0.8267	0.6275	0.7828	0.6089
Test	Before	0.7781	0.3567	0.7130	0.3690
Test	After	0.8089	0.5915	0.7141	0.5839

Since deep learning methods require high volume of data, therefore the performance of deep learning methods increases with the increase in the number of samples. In our study, since SMOTE increased the number of samples in the dataset by balancing the number of samples in each classes, thus the performance of the model after applying SMOTE has been increased and are highlighted in italic

ISSN: 1471-2105