Development of an Arabic HQAS-based ASAG to consider an ignored knowledge in misspelled multiple words short answers

Many traditional question answering systems depend on an Automatic Short Answer Grading (ASAG) to evaluate misspelled multiple words short answers in an Arabic Language using common edit-based algorithms such as Hamming, Levenshtein, and Jaro_Winkler, but they ignore and hide a big amount of a significant knowledge of the student answer. In this paper, we have implemented a proposed edit-based Hierarchical question answering system (HQAS) using a traversing by Breadth-First Search (BFS) within an m-ary tree to consider the ignored significant knowledge due to the misspelling at the middle of the dual-ordered incomplete answer, the misspelling at middle and the end of the intra-ordered incomplete answer, and the misspelling due to switching in words of the intra-ordered complete answer. It can differentiate among the students based on their significant hidden knowledge and show a distribution of knowledge content on different depths of the topic to determine which of the topic depths the student has the most significant knowledge.