Research On Priority Ranking Method Of Classes In Software System

With the rapid development of computer software technology,software computing model,application mode and product form are all undergoing tremendous changes.The software system structure becomes more and more complex.Software defects that will lead to unpredictable errors occur in the software system module will inevitably be hidden.However,when the errors occur,it's difficult for the existing defect location method to quickly locate the location of defects and the positioning accuracy is often not ideal.Therefore,if prevaluating the spread scope in the various class of potential defects in software system and classifying the influence of each class,you can focus on monitoring the higher vital class in software system and thus to access to the effective evaluation of software quality.Based on the above discussion,this paper also proposes a software network node in fluence sorting algorithm to evaluate the importance ranking of classes in software syste ms.This paper first abstracts the class of software source code as a node and abstracts the dependency between classes as edges.That's,it extracts the class and class relationships in the origin software code as software networks.Then,according to the transferring relationship of method among class,this paper defines a weighted software network and further propose IRMN(Importance Ranking Method of Nodes)based on the weighted software network nodes.By calculating the influence of nodes in the software network and sequence.And then evaluate the propagation effects of the potential defects of each class in the software system.In order to verify the effectiveness of the algorithm,this paper selected three open source software system source code for defect implantation experiments.Firstly,all the classes in the software source code are implanted with defects,and the spreads of the defects of each class are calculated in the whole software source code.Compared with the sort results of the proposed algorithm,the correctness of the algorithm is verified.Then,by contrasting with other algorithmic experiment,it is verified that the result of this algorithm has high accuracy.Furthermore,the further experiments show that external connecting affect more influence on the importance of the classes than the inner structure.