Optimized statistical analysis of software trustworthiness attributes

Software trustworthiness has become one of the key restrictions for software service quality and the development of the software industry. However, trustworthiness attributes interlace structured and dynamical coupling relations, which causes great barriers for trustworthiness measurements of large-scale software. According to the dynamical evolutionary characteristics of software trustworthiness attributes, this paper proposes a new approach for optimizing the trustworthiness measurement in terms the kernel trustworthiness attributes, and improves a downsize-optimized statistical analysis method for software trustworthiness attributes based on their nonlinear relations. The improved method considerably simplifies the trustworthiness assessment of largescale software. Using theoretical analysis and numerical simulations, the feasibility of this method is verified using two typical examples that illustrate the realization of the trustworthiness measurement.