A method of core concept extraction based on semantic‐weight ranking

Abstract

SummaryIn order to improve the efficiency of domain ontology concept extraction, this paper proposes a method of semantic‐weight ranking to extract core concepts. Firstly, we make segmentation processing for corpus texts and obtain the candidate concept set, and then analyze the correlation between the candidate concepts through the method of semantic similarity to form a mesh graph structure. We secondly calculate the semantic‐weight of the concept according to the generated mesh graph. Ultimately, we realize the core concept extraction via ranking different semantic‐weight. At the same time, we give the fuzzy relationship representation of the core concepts through the statistical characteristics of the core concepts' membership. On the foundation of the extraction method, we carried out the experiments with the financial text and had achieved an accuracy rate of 81.8%, which could effectively extract the core concepts and play a positive role in the content construction, knowledge sharing, and knowledge reuse.