Research Of Software Component Classification Based On Faceted Classification And Artificial Intelligence

The research of the retrieval method of software component is being a focus on component library researching. The method of this paper combines faceted classification and artificial intelligence that improves the efficiency and veracity when searching. It is given a system model of software component classification. Clustering analysis and self-organizing map that both are unsupervised leaning method of artificial intelligence are used in the retrieval method of software component.That the terms of faceted classification that are consisted of words, sentences or documents are uncontrolled words can reduce the subjective factors produced by persons. Such method can describe characters of component exactly. Moreover, it is advantageous to compute the similarity between components that the contents of component documents focus on one direction such as function document. The unsupervised leaning method of artificial intelligence can reduce the difficulty of the skills which user masters component library technique. It is also improved the efficiency of retrieval of component.Clustering analysis that is used in component classification according to semantic features of component document can improve the efficiency and veracity of component retrieval. It is tolerant for self-organizing map that information contains noisy factors.