个性化英语学习系统中的关键技术分析-教育技术学专业论文.docx

哈尔滨工业大学工学硕士学位论文 哈尔滨工业大学工学硕士学位论文 -II- -II- Abstract Nowadays it is necessary for people to be equipped with a good knowledge of English.English learning is an integrated process joining vocabulary memoriz ing, grammar rules learning and exercises of listening, and all these sections are interrelated. The ideal process of English should combine those sections as a whole rather than independent parts. Characteristics of English learning put forward a demand for assisted English learning system: How to build a multifunctional integrated English learning environment to allow users to learn English more conveniently. Meanwhile,with the continuous development of the Internet,the amount of English learning materials enriched constantly,so it becomes very important to help users to find English learning materials from vast amount of data which are suitable for themselves,it directly affects the learning time cost and learning interests of users. We put theories of natural language processing, information retrieving and machine learning into practice to satisfy the demands faced in personalized resource organization and query in English learning assisting system. Against the English learning problems mentioned above,this paper mainly includes the following aspects: Firstly,to organize resources effectively and provide data support to analyze the English level of users,we propose a classification method for English documents based on the degree of difficulty in English reading.Specifically speaking, we take ensemble learning framework as a basis,and train a machine learning model based on different feature spaces(grammatical features, unigram model, non -text feature) to improve classification results.We prove and analyze the effectiveness of our algorithm through adopting accuracy,recall and F-measure,we finally get an difficulty level discriminant model which is effectively used in practice. Secondly,to achieve the purpose of organizing resources based on