Classical Chinese plays an important role of inheriting traditional culture, but many high school students are not able to learn it well. Mastering the frequently used notional words is the key to improve the reading ability of Classical Chinese. High-quality testing questions are helpful to test and solidate students' mastery of Classical Chinese notional words. However, conventional manual question generation methods are time-consuming, labor-consuming and subjective. With the blossoming of educational intelligence, this paper studies the automatic question generation(AQG) of Classical Chinese notional words frequently used in high school.

Firstly, this paper discusses the overall strategy of AQG, including question generation goal, question generation form and question generation strategy. With regard to question generation goal, this paper summarizes the text features of Easy Classical Chinese, determines the quantity and types of Classical Chinese notional words frequently used in high school, discusses the vocabulary knowledge of Classical Chinese notional words, and transforms the vocabulary knowledge into testing points. With regard to question generation form, this paper designs question types covering a variety of language information and determines the principles of question generation and option generation. With regard to question generation strategy, a question generation framework is established, which is composed of vocabulary knowledge base generation module, question generation resource generation module, automatic question generation module, difficulty calculation module and question evaluation module.

Starting from the strategy of AQG, a set of basic resources for AQG of Classical Chinese notional words are constructed. At first, this paper analyzes and extracts the attributes of frequently used Classical Chinese notional words, such as word pronunciation, part of speech, lexical forms, similar words, frequency, word level and word senses, and finally builds a vocabulary knowledge base. Secondly, this paper integrates the resources of Chinese Textbooks for high school, college entrance examination questions and ancient classics to get sentences that used to generate questions, carrys out word meaning annotation, and finally builds an AQG dataset with 55400 labeled data and more than 1.25 million words. This dataset is used in AQG and Classical Chinese word disambiguation. In addition, this paper also collects the different flexible uses of part of speech, words with the same meaning in ancient and modern times, words with the different meaning in ancient and modern times, interchangeable words, words with variant pronunciation and other resources.