Approach For Measuring Software Behavior Model Based On FA

The paper researches an approach of measuring software behavior model to compare the inference techniques of the model by quantifying the differences between software behavior models. Software behavior model plays an important role in the area of software security, software behavior models are used more and more in specific applications rather than only theory, and inference techniques are invented with the development of the computer's performance. The topics like an inference technique's measuring and the technique's performance become of significance and of value.From the view of engineering, the best way of measuring a mode of production is measuring and comparing its products. Accordingly, for the purpose of measuring the inference technique of software behavior model, the paper measures the software behavior model firstly and selects the most representative software behavior model which is FSM. The method of measuring FSM combined the results of topology analysis and language-set analysis. The technique of topology analysis shows the differences of the two FSM compared in the form of set and the technique of language-set analysis empowers the local elements of a FSM. The combination of the two results demonstrates the priorities of different local elements of a FSM with the extended expression of Precision and Recall which usually used to compare two objects. The paper used the extended expression to measuring software behavior models and the inference techniques of the models.In this paper, the study of topology analysis referenced the cognitive process a human might undertake when navigating through an unfamiliar landscape with a map to measuring software behavior models, and quantified the measuring results. The study of language-set analysis used depth-first algorithm to achieve the language set of a FSM, and proposed a new strategies of empowering the elements of a FSM. The method combined the results of topology analysis and language-set analysis expanded the research on measuring software models. By applying the two algorithms and measuring two software behavior models and their inference techniques, the paper testified the feasibility of the combined approach and laid foundation on constructing evaluation system for further research.