A Catalog of Techniques that Predict Information about the Count or Rate of Field Defects

The Gamma modeling technique is a finite SRGM-based modeling technique. Prior work uses this technique to fit a SRGM based on the Gamma function using software process metrics that measure development defects and then uses the fitted model to make predictions. The cost of use of this technique is lower than typical. Overview Inputs • The occurrence time of each development defect or the defect count in each time interval during development Outputs • The predicted field defect rate Model This technique adjusts the and model parameters of the SRGM based on the Gamma function so that the SRGM describes the observed development defect information. Two mathematically equivalent forms of the SRGM, which are used to describe the field defect rate and the field defect count, are: Field defect rate (for the t-th time interval) = t te β αβ − 2 , and Field defect count (aggregated from time 0 to time t) = ) ) 1 ( 1 ( t e t β β α − + − . The parameter roughly determines the scale of the model, and the parameter roughly determines the shape of the model. Gamma functions, which are used to predict the field defect rate, and Gamma cumulative functions, which are used to predict the field defect count, with sample parameter values are in figures 1 and 2.