The research of the entity relation extraction based on statistical model

Abstract

Entity relation extraction is the important research field in many languages processing, such as the knowledge management, semantic network, information retrieval and information extraction. In this paper, some specific entity relations can be extracted in the football match domain, firstly, the definition of the five types relation had been defined. Based on the conditional random fields, aiming to each relation type, the different features were discussed. The approach is effective when evaluated on the data set. © 2010 IEEE.