文章基本信息

标题：IMPROVING ARABIC QUESTION ANSWERING SYSTEM BY MERGING ANER TECHNIQUE, UPDATED QUESTION CLASSIFICATION TECHNIQUE AND STOP WORDS TECHNIQUE
本地全文：下载
作者：BELAL ABUATA ; IYAD ALAZZAM ; FADI ALKHAZALEH 等
期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2020
卷号：98
期号：23
页码：3757-3769
出版社：Journal of Theoretical and Applied
摘要：Question answering systems are basically of three types: open domain, closed domain and restricted domain. In this paper we build a question answering system that tries to take users' questions, and provide them the answer in Arabic language. The proposed system is called AMAU (Arabic question answering system by merging Aner technique, updated question classification technique and stop words technique). AMAU focuses on question processing and document processing. After building the AMAU system, it was evaluated based on precision, recall and f-measure by ANER QA and ANER data set. The AMAU system is also compared with Arabic NER system using precision, recall and f-measure. The evaluation results show a better performance by AMAU. AMAU produced an average precision of up to 80% compared to Arabic NER system that produced an average precision of up to 63%. The proposed system performed best when stemming and stop words were used and resulted in the highest values of Recall, Precision and F-measure compared to NER system.
关键词：Information retrieval;Question Answering;Hybrid Approach