Research on Component Retrieval and Matching Methods

Abstract

Aiming at how to quickly retrieve the target component from the huge component library, this paper proposes a component retrieval method based on facet classification, which uses component keyword set to facet description of the component and constructs the component vectorized model. The idea of layering is adopted to divide the components in detail. The K-means clustering algorithm based on the maximum and minimum distances is used to cluster the components, so that the problem of component matching is transformed into the calculation of the cosine similarity of component vectors. This method can narrow down the search scope and effectively improve the search efficiency.