基于标准结构熵偏差率的软件网络度量研究

基于软件静态网络宏观拓扑特征值的特性,提出了软件网络特征值偏差率以及软件相对规模质量特征和结构有序性的概念.并使用标准结构熵偏差率来衡量软件网络特征值与目前被广泛使用且规模相同的开源软件的网络特征值平均值的差距,以此来反映特征值对相应软件的相对规模有序性的影响.最后采用单样本K-S检验方法对标准结构熵偏差率进行验证,并得出结论:当标准结构熵偏差率大于0.05时,软件在系统设计以及代码编写方面均存在着一定的问题,软件的相对规模结构有序性存在较大的提高空间.

Research of Software Network Measurement Based on The Deviation of Standard Entropy

Based on the characteristics of the macro-topology eigenvalues of a software network, the deviation ratio of eigenvalues of a software network, the software quality relative to its scale and the structural orderliness of software are defined conceptually. With the difference between the eigenvalues of software network and the mean value of eigenvalues of the open-source software which is widely used at present on the same scale was measured by the deviation ratio of standard entropy, how the eigenvalues affect the orderliness relative to the scale of relevant software are revealed. Verifying the deviation ratio of standard entropy via a single-sample K-S test, a conclusion is drawn that when the deviation of standard entropy is greater than 0.05, the software has certain problems in system design and encoding, and there is plenty of room for improvement in the orderliness relative to the structural scale of software.