A Parse Tree Based Computation Technique for Generating Comprehension Style Questions from Chinese Text

This paper describes our research and development work on a computational method that takes a piece of Chinese unstructured text and generates a set of questions and answers as the output. Our method is largely based on Heilman & Smith’s over-generation approach [1] and is included with techniques specific for handling Chinese text. Using the syntactic and semantic features identified in a sentence, various question types can be generated with answers also available. Automatic question generation is potentially a key component in future intelligent e-learning systems, but it is also a very challenging problem. A major objective of this work is to investigate technical issues and limitations that would provide direction of future research.