Entity and relation extraction from Arabic text using morphological analysis.

Abstract

Natural Language Processing concerns the automatic understanding of natural languages, such as entity and relational extraction from text documents. Morphological analysis is key for Arabic natural language processing since Arabic is a morphologically rich language. Researchers used techniques that work well for the Latin based languages such as local grammars, statistical learning models, pattern matching, and rule- based techniques to automatically extract named entities from Arabic text. These techniques need to boost their results by using application specific corpora, parallel language corpora, and morphological stemming analysis. We present a method for extracting entities and events, from Arabic text using a hierarchy of finite state machines driven by morphological features such as part of speech and gloss tags. We also use graph transformation algorithms and cross-document analysis to enhance entity extraction and extract relations amongst entities. We evaluated our method on three natural language processing applications. (1) We automated the extraction of temporal entities. (2) We automated the extraction of narrators, narrator relations and corresponding biographies from several corpora of Islamic hadith and biography books. (3) We automated the extraction of genealogical family trees from Biblical texts. In all applications, our method reports high precision and recall and learns lemmas about phrases that improve results.