| Summary: | 碩士 === 國立中正大學 === 資訊工程研究所 === 90 === Due to the advent of Internet, World Wide Web becomes a new popular medium for education, including distance learning, multimedia courseware authoring, and online adaptive test. Online adaptive test is an assistant learning tool that .examine the ability of students and get items from the item bank. There are many various domain in course such that a lot of time and human effort to find classified items is needed. In this thesis, a Chinese item classification model is proposed. In our classification model, one item is represented by serveral kinds of features such as information about positions of symbols/ key word, the length of symbols/ keyword ,term frequency and the degree of centralized. Then we use the learning capability of artificial backpropagation neural network to apply to item classification model. The experimental results indicate that the item feature vector learning is an efficient classification model in Chinese item.
|
|---|
