A Computational Method of Complexity of Questions on Contents of English Sentences and its Evaluation

This paper describes a computational method of complexity of a question for an adaptive question and answer function in an intelligent support system for English learning, and its evaluation. To realize adaptive question and answer, systems should generate questions depending on both educational intentions and learner's understanding state. For generating suitable questions for a learner automatically, systems must know the factors which influence difficulty of questions, and have ability to calculate difficulty. Difficulty is composed of a learner dependent part and an independent part. The former is evaluated by referring to a student model. The latter is defined by enumerating factors which influence complexity of questions. We present a definition of complexity of questions along learners' answering process. We also describe experimentation of comparing the complexity of questions calculated by computer with the complexity evaluated by human. 1. Introduction QA (Question and Answer) about the contents of a story is widely used in language learning. QA in a target language is effective for acquiring practical skills because learners use plural language skills to answer questions, i.e. to grasp the contents of a story and question sentences and to compose answer sentences. When a teacher and a learner practice such QA, the teacher will give suitable questions for the learner. That is to say, the teacher will give an easier question to the learner who is doing badly and give a more difficult question to the learner who is doing well. Some computer assisted language learning systems are equipped with test functions which ask about the contents of a story. Most of them, however, use questions prepared beforehand [4, 9], so teachers are burdened with preparing questions and answers for every learning material. And it is very difficult to prepare sufficient amount of questions and answers to correspond with various states of learners' understanding. The target of our study is to realize a QA function which provides suitable questions for each learner and tailored advice according to the learner's answers. To achieve the target , we need the following functions: (1) to understand En-glish sentences, (2) to generate various kinds of question sentences automatically, (3) to select suitable questions for each learner from a set of generated question sentences and (4) to analyze learners' answer sentences and to diagnose errors. In the previous studies, we have proposed a method of analyzing stories and representing their meanings [7] for the function (1), …