Research on Matrix Grade Evaluation Based on Design Pattern Detection
XIAO Zhuo-yu1，LI Yan2，HE Pei3，CHEN Jun-xu1
1（Foreign College of Central South University of Forestry and Technology，Changsha 410200，China）
2（Hunan Highways Administration，Changsha 410209，China）
3（School of Computer Science & Education Software Guangzhou University，Guangzhou 510006，China）
Abstract：Design pattern identification is an important element of many solutions of Software Reuse.However，detection of design pattern instances is very important and helpful for software understanding and maintenance.Hence，in order to solve the matching accuracy problem between software system and design pattern，reducing the complexity of the identification，This paper presents a novel method for design pattern detection based on matrix grade evaluation of the difference between the vertex.The formalization of matrix software system and design pattern representation，Modification of the preliminary information based on decomposition class diagrams generated by reverse engineering algorithm is designed.Then，the results will be translated into a graph based representation of the adjacency matrix.Calculated in accordance with the rules of matrix grade，it will be transformed into MatrixGrade representation result based matrix，Finally，evaluation on opensource projects demonstrated the accuracy and the efficiency of the proposed method，which is compared to other wellknown algorithms，indicated that this algorithm more research of using matrix grade evaluation is needed.
肖卓宇,黎妍,何锫,陈俊旭. 基于矩阵积分评估的设计模式检测研究[J]. 小型微型计算机系统, 2016, 37(7): 1428-1433.
XIAO Zhuo-yu,LI Yan,HE Pei,CHEN Jun-xu. Research on Matrix Grade Evaluation Based on Design Pattern Detection. Journal of Chinese Computer Systems, 2016, 37(7): 1428-1433.
2016, Vol. 37 
