R2 measures the power of correlation between predicted and actual number of
faults [37]. Another term for this parameter is the coefficient of multiple
determination, and this parameter is widely used in studies that predict the number
of faults. Many researchers have applied this parameter in their studies [9, 36, 8,
37, 31, 3]. This metric’s value should be near to 1 if the model is to be acceptable,
and Formula 22 is used to calculate this parameter. The actual number of faults is
represented by yi, yˆi represents the predicted number of faults, and y shows the
average of fault numbers.