Defect Prediction Technology of Aerospace Software Based on Deep Neural Network and Process Measurement

Table 6

Statistical results of experimental data.

Software project	Experimental steps	Accuracy	Recall	Precision	F1-score
XX flight control software	A1-2	0.8843	0.7358	0.6814	0.6767
	A1-3	0.8256	0.7096	0.6722	0.6685
	A2-3	0.8911	0.7504	0.7264	0.7142
	A1-2-3	0.9106	0.7826	0.7581	0.7451
XX fire control software	A1-2	0.9372	0.8025	0.7633	0.7562
	A1-3	0.9180	0.7634	0.7052	0.6882
	A2-3	0.9477	0.8175	0.7721	0.7580
	A1-2-3	0.9541	0.8260	0.7891	0.7726
XX accused software	A1-2	0.8859	0.7246	0.6925	0.6827
	A1-3	0.8663	0.7044	0.6527	0.6475
	A2-3	0.8802	0.7330	0.6992	0.6833
	A1-2-3	0.8979	0.7421	0.7143	0.7007