作者: I. Ocampo-Guzman , I. Lopez-Arevalo , V. Sosa-Sosa
DOI: 10.1109/ICEEE.2009.5393402
关键词:
摘要: In this paper an approach to construct ontologies based on a text corpus is described. Using Latent Dirichlet Allocation the topics that describe documents contained in are identified. Each topic formed by set of terms whose semantic relatednesses determined applying distributional hypothesis, which considers as similar those share linguistic context. This context described verbs they share. The concept each topic's modeled through taxonomy describes relation between them.