NONLINEAR REGRESSION MODELS FOR ESTIMATING THE DURATION OF SOFTWARE DEVELOPMENT IN JAVA FOR PC BASED ON THE 2021 ISBSG DATA

Context. The problem of estimating the duration software development in Java for personal computers (PC) is important because, first, failed often main contributor to projects, second, a popular language, and, third, computer widespread multi-purpose computer. object study process PC. subject nonlinear regression models estimate PC.
 Objective. goal work build PC based on normalizing transformations and deleting outliers data increase confidence estimation comparison ISBSG model platform.
 Method. models, confidence, prediction intervals regressions are constructed non-Gaussian with help appropriate techniques. techniques transformations. Also, we apply outlier removal construction. In general, above leads reduction mean magnitude relative error, widths without application construction process.
 Results. A decimal logarithm transformation Johnson (for SB family) Box-Cox as both univariate bivariate ones has been performed.
 Conclusions. transformation. This model, other smaller effort values that bigger than 900 person-hours. prospects further research may include sets construct languages platforms, example, mainframe.