Balanced Adversarial Tight Matching for Cross-Project Defect Prediction

Table 2

Twenty handcrafted features.

Metric	Description
WMC	This measures the average number of weighted methods in a class [34]
DIT	It represents the number of direct and indirect parent classes [34]
NOC	This feature counts the number of direct subclasses a class has [34]
CBO	It indicates the level of association between a class and other classes [34]
RFC	It represents the total number of methods a class can respond to [34]
LCOM	It indicates the level of relatedness and sharing among methods in a class [34]
LCOM3	An enhanced version of LCOM that calculates the sharing of member variables among methods more accurately [34]
NPM	This counts the total number of public methods in a class [35]
LOC	It measures the total number of lines of code in a class [35]
DAM	It measures the degree to which a class depends on external classes [35]
MOA	It measures the degree to which a class aggregates (contains) other classes [35]
MFA	It measures the proportion of abstract methods in a class [35]
CAM	It measures the level of cohesion among methods in a class [35]
IC	It measures the degree to which a class depends on inheritance [33]
CBM	It measures the level of method invocation between methods in a class [33]
AMC	It represents the average complexity of methods in a class [33]
CA	It represents the number of classes referencing a class [36]
CE	It represents the number of classes referenced by a class [36]
Max CC	It represents the highest complexity of a method within a class [37]
Avg CC	It represents the average complexity of all methods in a class [37]