Table 8 Performance improvements across different datasets showing project-specific characteristics.

From: An integrated graph neural network model for joint software defect prediction and code quality assessment

Dataset

File count

Complexity

Baseline F1

Our F1

Improvement

Our AUC

AUC gain

Apache Ant

1,248

Low

0.755

0.811

+ 7.4%

0.889

+ 6.8%

Eclipse JDT

2,156

Medium

0.741

0.808

+ 9.0%

0.892

+ 7.2%

Apache Camel

3,874

High

0.729

0.811

+ 11.2%

0.901

+ 8.5%

Mozilla Firefox

4,521

High

0.738

0.815

+ 10.4%

0.898

+ 8.1%

Apache Hadoop

2,893

Medium

0.746

0.809

+ 8.4%

0.894

+ 7.4%

Linux Kernel

8,456

Very High

0.732

0.811

+ 10.8%

0.899

+ 8.3%

Back to article page