Abstract:

Aiming at the problem that non marking software metric metadata was easy to fall into the local optimum of the K-means process in the spectral clustering algorithm, we proposed a new algorithm of chaotic immune clustering algorithm. Under the framework of the immune clonal selection algorithm, the method of antibody affinity calculation was designed for the evaluation of immune clonal clustering center, and the hierarchical chaotic mutation operator was given, which further improved  prediction performance of non marking software metric metadata. Simulation results show that the algorithm is effective.

Key words: software defect prediction, software metric, chaos theory, spectral clustering