TY - JOUR
AU - І. Kulakovska
PY - 2019/12/28
Y2 - 2021/12/02
TI - The non-linear regression equations for software size assessment with java mp3players type.
JF - COMPUTER-INTEGRATED TECHNOLOGIES: EDUCATION, SCIENCE, PRODUCTION
JA - cit
VL -
IS - 37
SE - Automation and Control
DO - 10.36910/6775-2524-0560-2019-37-11
UR - http://cit-journal.com.ua/index.php/cit/article/view/42
AB - The article discusses different approaches to assessing the complexity of software development (software). The dependence of the complexity of the software development on the size of the project is analyzed. The basic types of existing metrics and the possibility of their application are described. Metrics are offered to compare metrics. A nonlinear regression model for estimating the size of open source MP3players software on JAVA is based on the normalization of a six-dimensional non-Gaussian dataset (actual program size in thousands of lines of code, total classes, total links, etc.) in the conceptual data model of 32 programs using a nonlinear regression equation. The model constructed, compared to other regression models (both linear and nonlinear), has a larger multiple coefficient of determination, less than the mean value of the relative error. For the obtained equation, the normal distribution of residuals was investigated, and the confidence interval widths for nonlinear regression were calculated.
ER -