The neural network model of communication software maintainability assessment

In order to assess the communication software maintainability in a credible, simple and shortcut way, The fuzzy comprehensive assessment system based on BP neural network is used in this paper. By introducing the fuzzy matrix and BP neural network, the assessment conclusion is accurate when the model is trained. The model mainly consists four steps: the first one is to construct the index system of the software system. The second one is to construct the basic data model. In this step, we need to confirm the fuzzy evaluation matrix, the comprehensive assessment matrix, the evaluation index´s weight and the comprehensive assessment result. The third one is to construct the fuzzy comprehensive assessment system based on BP neural network. In this step, expert review samples is used as the learning samples of neural network and then assessment system is used to get the assessment conclusion by calculation. The last one is to use expert evaluation system based on three level BP neural network. After test, it´s appoved that there are the application value of using the fuzzy comprehensive assessment system based on BP neural network to Communication software maintainability´s assessment.