Estimating Quality of Open-Source Kotlin-Based Apps by the Confidence and Prediction Intervals of Nonlinear Regression for RFC Metric

Presently, there is an increasing preference for open-source software in Kotlin, which is being utilized both as a software tool and as a component for apps. As a result, the selection of high-quality open-source software holds significant importance. In contemporary times, several software metrics, such as the response for a class (RFC) in the app layer or average RFC, are employed to assess the quality of software. However, there are no models and methods based on the software metrics to determine the quality of open-source apps developed in Kotlin. To estimate the quality of open-source apps developed in Kotlin, we applied a technique for detecting software quality based on the prediction and confidence intervals of nonlinear regression for the average RFC depending on the average CBO (coupling between object classes) and average WMC (weighted methods per class). We applied the values of the above metrics from 52 open-source apps developed in Kotlin and hosted on GitHub to build the prediction and confidence intervals of nonlinear regression for average RFC using the three-variate Box-Cox transformation.