Scientific Programming / 2019 / Article / Tab 7

Research Article

Local versus Global Models for Just-In-Time Software Defect Prediction

Table 7

Local vs. global models in the timewise-cross-validation scenario.

IndicatorsProjectGlobalLocal (k = 2)

AUCBUG0.672 ± 0.1250.545 ± 0.124
COL0.717 ± 0.0750.636 ± 0.135
JDT0.708 ± 0.0430.645 ± 0.073
MOZ0.749 ± 0.0340.648 ± 0.128
PLA0.709 ± 0.0540.642 ± 0.076
POS0.743 ± 0.0720.702 ± 0.095
W/D/L0/0/6

F1BUG0.587 ± 0.1080.515 ± 0.128
COL0.651 ± 0.0660.599 ± 0.125
JDT0.722 ± 0.0480.623 ± 0.161
MOZ0.804 ± 0.0500.696 ± 0.190
PLA0.702 ± 0.0530.618 ± 0.162
POS0.714 ± 0.0650.657 ± 0.128
W/D/L0/0/6

ACCBUG0.357 ± 0.2120.242 ± 0.210
COL0.426 ± 0.1680.321 ± 0.186
JDT0.356 ± 0.1480.329 ± 0.189
MOZ0.218 ± 0.1320.242 ± 0.125
PLA0.358 ± 0.1810.371 ± 0.196
POS0.345 ± 0.1590.306 ± 0.205
W/D/L0/4/2

BUG0.622 ± 0.1940.458 ± 0.189
COL0.669 ± 0.1500.553 ± 0.204
JDT0.654 ± 0.1010.624 ± 0.142
MOZ0.537 ± 0.1100.537 ± 0.122
PLA0.631 ± 0.1150.645 ± 0.151
POS0.615 ± 0.1310.583 ± 0.167
W/D/L0/4/2

Article of the Year Award: Outstanding research contributions of 2020, as selected by our Chief Editors. Read the winning articles.