ABSTRACT
The need of questions and answers is prompted for various purposes, e.g. self-study, academic assessment, and coursework. However, the conventional way to create question-answer pairs has been both tedious and time-consuming. In the present study the authors propose an automatic question generation for sentences from text passages in reading comprehension. Authors introduce a rule-based automatic question generation for the task, as well as implement statistical sentence selection and various configurations of named entity recognition. Three types of WH-questions (“What”, “Who”, and “Where”) can be produced by this system. The system performs well on generating questions from simple sentences, but weakens on more complex sentences due to incomplete transformation rules.
