Automatic question setting method about query type short answer question of English reading comprehension test

Abstract

The invention discloses an automatic question setting method about a query type short answer question of an English reading comprehension test. The automatic question setting method comprises the four steps of natural language processing, testing centre selecting, interrogative sentence generating and synonymy rewriting. According to the automatic question setting method, natural language processing is conducted on an input article; then a testing center sentence is selected according to word frequency intensity, paragraph length and sentence meaning similarity; a declarative sentence is converted into an interrogative sentence according to the principle of Lexical-Functional Grammar; finally, synonym replacement and pronoun replacement are conducted on the interrogative sentence, and the query type short answer question is formed. According to the automatic question setting method, due to the fact that the testing center selecting and synonymy rewriting are added, the generated interrogative sentence can be applied to the reading comprehension test, due to the fact that the methods of limiting the vocabulary range and the sense range are adopted in the synonymy rewriting part, the bottle neck of low semantic disambiguation precision can be broken through, and synonym replacement with a high accuracy rate is achieved. Due to the fact that an interrogative sentence part refers to syntax and semantic information, interrogative sentences with various types and high quality can be effectively generated.

Description

technical field [0001] The invention belongs to the technical field of automatic language testing, and in particular relates to an automatic proposition method for interrogative short-answer questions in English reading comprehension tests. Background technique [0002] Computerization is an important development direction of modern educational testing. At present, the language test has been able to realize computer adaptive testing in the testing link, and automatic machine scoring of subjective questions in the scoring link. However, in the proposition link, the level of automation is still very low, and test takers basically only use computers for text editing and dictionary lookup. auxiliary. [0003] There is an urgency to computerize the link of language test proposition. In the development of standardized reading comprehension tests, the cost of artificial propositions is high, but the efficiency is relatively low. Proposers need to receive professional training, a...

AI Technical Summary

Problems solved by technology

Although this method considers the semantic relationship, because the named entity recognition method for proper nouns is used to distinguish lexical categories, common nouns cannot be used as the object of questioning.

[0008] Xuchen Yao, GosseBouma and Yi Zhang proposed a method of generating question sentences based on the minimal recursive semantic theory in the paper titled Semantics-based question generation and implementation (Dialogue&Discourse, 2012, 11-42). This method utilizes deep semantic structure and can Generate high-quality questions, but this method is computationally intensive and inefficient

[0009] Questions generated by the above methods are not suitable for reading comprehension tests because they are not synonymously rephrased, and these methods do not include a mechanism for screening questions for the test

Images