A WordNet-driven approach to vegetable supply chain domain concepts acquisition

Abstract

Domain concepts acquisition is a key step of knowledge acquisition. In this paper we propose a new method for computing the relationship of domain concepts of China's vegetable supply chain using WordNet and Locally Linear Embedding (LLE) Arithmetic. WordNet is used to construct the high dimension description of the domain concepts and LLE is used to compute the relationship of concepts in low dimension space. The concepts description in the high dimension is in accord with human cognition and the low dimension relationship can reflect the degree of correlation of domain concepts clearly. A computer can deal with domain concepts directly after dimensional transformation. The experiment proves the efficiency of this method. The method can be expanded to deal with other examples of domain concepts acquisition. © 2007 Taylor & Francis Group, LLC.