Study on Concept Adaptive Extraction of Agricultural Domain Based on Bayesian Network

This paper proposes a method of agricultural concept extraction with adaptivity, in order to elevate the quality of automatic or semi-automatic construction of agricultural ontology. This method constructs a model of Bayesian network combining context dependency analysis, domain dependencies and mutual information; then achieves conditional probability distribution table by means of data training. In the concept extraction process, by analyzing on precision and recall of extracted concepts, and Bayesian being reasoned backward with prior knowledge of conditional probability distribution table, does it been positioned the threshold to be adjusted , eventually achieves adaptive extraction of agricultural concept.