An Improved COSMIC Software Evaluation Method

Abstract

Under the background of the Internet era, there are many kinds of software applicable to various business scenarios everywhere. How to evaluate the scale of a software is not only related to the investment of software development time, but also affects the investment of capital and personnel in software development. At present, the mainstream software scale evaluation method is the full Function Point analysis (COSMIC-FFP) proposed by the International Association for Universal Software Measurement. COSMIC software evaluation method through the retrieval of all the data movement process, each data movement process as a function point (CFP), and finally through the statistics of the number of CFP to get the size of the entire software measurement this paper proposes an improved COSMIC software evaluation method: according to the software needs description, grammar and syntax. For the relatively fixed language features such as parts of speech, we should develop corresponding segmentation tagging and error filtering rules, introduce natural language processing ideas to process software function points, and build local rule base to weaken the over-dependence of appraisers on rule guide and reduce the difficulty of getting started. Experiments and actual evaluation show that the improved method in this paper not only improves the accuracy, but also improves the time efficiency more obviously.