TY - GEN
T1 - Software behavior model measuring approach of combining structural analysis and language set
AU - Xue, Jingfeng
AU - Zhang, Yan
AU - Hu, Changzhen
AU - Ren, Hongyu
AU - Li, Zhiqiang
N1 - Publisher Copyright:
© Springer International Publishing Switzerland 2015.
PY - 2015
Y1 - 2015
N2 - Structural analysis represented by FSMDiff algorithm is the main measuring approach for existing software behavior model which is based on finite state automata. This method just focus on the data structure of finite state automata as figure characteristics, however, as software behavior model, it is more important for finite state automaton to reflect the characteristics of software behavior. So we need to find out a method to distinguish the importance in the finite state automata between different state nodes. This paper shows how the output of the FSMDiff algorithm can provide a quantified expression of structural difference between two models. According to this, we also introduce the language-set analysis, which uses the depth-first traversal algorithm to solve the language set of finite state automata. Above all, we propose a new strategy of assigning weights for the local elements of software behavior model, which can fusion assigning weight results and structural analysis for evaluation of software behavioral models. Experiment results demonstrate the effectiveness and feasibility of software behavioral model measuring approach of combining structural analysis and language set, and laid the foundation for constructing evaluation system of software behavior model inference technology.
AB - Structural analysis represented by FSMDiff algorithm is the main measuring approach for existing software behavior model which is based on finite state automata. This method just focus on the data structure of finite state automata as figure characteristics, however, as software behavior model, it is more important for finite state automaton to reflect the characteristics of software behavior. So we need to find out a method to distinguish the importance in the finite state automata between different state nodes. This paper shows how the output of the FSMDiff algorithm can provide a quantified expression of structural difference between two models. According to this, we also introduce the language-set analysis, which uses the depth-first traversal algorithm to solve the language set of finite state automata. Above all, we propose a new strategy of assigning weights for the local elements of software behavior model, which can fusion assigning weight results and structural analysis for evaluation of software behavioral models. Experiment results demonstrate the effectiveness and feasibility of software behavioral model measuring approach of combining structural analysis and language set, and laid the foundation for constructing evaluation system of software behavior model inference technology.
KW - FSMDiff algorithm
KW - Finite state automata
KW - Language-set
KW - Software behavior model
KW - Structural analysis
UR - http://www.scopus.com/inward/record.url?scp=84958035315&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-27998-5_9
DO - 10.1007/978-3-319-27998-5_9
M3 - Conference contribution
AN - SCOPUS:84958035315
SN - 9783319279978
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 137
EP - 150
BT - Trusted Systems - 6th International Conference, INTRUST 2014, Revised Selected Papers
A2 - Yung, Moti
A2 - Zhu, Liehuang
A2 - Yang, Yanjiang
PB - Springer Verlag
T2 - 6th International Conference on Trusted Systems, INTRUST 2014
Y2 - 16 December 2014 through 17 December 2014
ER -