I Introduction
Entity definitions can be obtained to consist of dictionaries and domain glossaries for supporting the knowledge-based applications. However, manually extracting definitions requires the cooperative effort of many experts from multiple fields, which is laborious and time consuming. Furthermore, it is not feasible to manually extract the new definitions which are emerging constantly in the Web. Hence, automatically mining the complete and latest definitions from Web corpora is a significant task for constructing and updating glossaries. Automatic definition extraction is a useful tool not only in dictionary construction [1], but in other domains such as E-Learning [2], [3], question answering [4] and ontology engineering [5]–[7], taxonomy learning [8], [9], semantic predicate [10], etc. Currently, the automatic extraction of definitions from textual data has become a common research topic in several domains of Natural Language Processing (NLP) [11].