A statistical approach for universal networking language-based relation extraction

In the effort to enrich available information with machine-processable semantics, Universal Networking Language (UNL) was defined as an artificial intelligent language that is able to represent information and knowledge described in natural languages. One of the main components of UNL is a set of binary relations that represents semantic relationships between concepts in sentences. To provide machine-processable semantics for computers, extraction of such semantic relationships from natural language text is a must. In this paper, we present a method to solve the problem of UNL semantic relation extraction in English sentences. With the assumption that the positions of phrases in a sentence between which there exists a relation have been identified, we focus on the problem of classifying the relation between the given phrases. The UNL relation classifier was developed by using statistical techniques applied on several lexical and syntactic features. In addition to the common used features, we also propose a new feature that reflects the actual semantic relation of two phrases independent on words in the between. Using our new feature in this problem gives the preliminary results that have shown the promising advantages of the feature in some other semantic relation recognition tasks. The evaluation on dataset supplied by UNDL organization shows that our system obtained the result at about 79% accuracy.