Table 7. Evaluation results of software metrics-based models according Ferenc et al.’s (2019) dataset using only static metrics.

Approach	Precision (%)	Recall (%)	f1-score (%)
DNNs	87.34	59.96	71.11
DNNc	91.06	57.89	70.78
RF	93.11	57.82	71.34
KNN	90.88	65.91	76.40
Linear regression	84.31	15.44	26.10
Logistic regression	75.30	21.19	33.07
SVM	95.29	51.40	66.78
Tree	73.66	69.72	71.63
Bayes	22.38	11.70	15.36
Our stacking CNN classifiers	87.52	80.57	83.55
Our Stacking CNN classifiers(RUS)	81.59	81.48	81.46
Our stacking CNN classifiers(ROS)	94.46	94.44	94.44

Notes.

The bold values are the best results among other classifiers.