Estimation of program reverse semantic traceability influence at program reliability with assistance of object-oriented metrics

In article the approach to the estimation of program reverse semantic traceability (RST) influence on program reliability with assistance of object-oriented metrics is proposed. In order to estimate reasonability of RST usage it is naturally to define how it influences on major adjectives of software development project: project cost, quality of the developed application, etc. At present object-oriented metrics of Chidamber and Kemerer are widely used for predictive estimation of software reliability at early stage of life cycle. In number of works, for example, it is proposed to use logistic regression for estimation of probability π (that a module will have a fault). The parameters of this model are found by maximal likelihood method with calculation of object-oriented metrics. The paper shows how to change the software reliability model parameters, that was received using logistic regression, in order to estimate influence of program RST on program reliability.