Building a system based on natural languages processing to automatic question generation from arabic texts

This paper aims to describe the called Arabic Questions Generation (AQG) system. It automates the process of generating questions fill in the blanks from Arabic texts, based on pre-generated corpus pattern. The system as well uses Stanford Arabic Natural Languages Processing (NLP) tools to generate a morphological tagged tree from the Arabic text, which then be matched with the patterns to form the question. The system provides the teacher with a Graphical user interface to facilitate the process of generating a test from the set of Questions, as it has many complex problems and issues in the Natural Languages Processing. The system useful for teachers for the generating questions automatically Instead of the manual method. It used to assessing learners’ achievements of learning. Grades were calculated based only on generated questions from The System Out of all grading criteria, The System less in the syntactic correctness of generated questions. The System did the best in Relevance.