Research on Tracking Relationship Establishing Method of Train Control System Requirements Specification Based on Labeled-LDA

Abstract

The system requirements specification is the basis for the design and development of the train control system, and the test cases are compiled according to system requirements specification, which are of great significance to the function test of the train control system. Establishing the tracking relationship between train control system requirements specification and test cases is an important part of achieving completeness verification of test cases and requirements change management. In order to solve the problem that the train control system specification texts contain a large number of terms in the field, this paper proposes a strategy that combines the train control system terms dictionary with the conditional random field model to extract the train control system terms. By analyzing the features of the relationship between the system requirements specification and the test cases, we select the Labeled Latent Dirichlet Allocation (Labeled-LDA) model to model the specification texts, and determine the requirements tracking relationship by calculating the similarity between the requirements and the test cases. It is proved that the corresponding relationship between the train control system requirements specification and the test cases obtained by the above method is basically correct, and the precision and recall are high. The results show that this method can realize the purpose of automatically establishing the system requirements specification tracking relationship.