Integration of Heterogeneous Itembanks for E-Learning

• Main Authors: Shih-Pang Tseng, 曾世邦
• Other Authors: Chu-Sing Yang
• Format: Others
• Language: en_US
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/79195697390722169037

博士 === 國立中山大學 === 資訊工程學系研究所 === 101 === Learning is a major human activity to better accommodate the environment or the society. The e-learning was proposed and developed in the past twenty years to enhance the human learning efficiency. Many information and communication technologies, like multimedia and web, have already applied on the e-learning successfully. The personalization of learning can provide different learning contents and activities to different learners with various personal characteristics. To accomplish the personalization of learning, it is an important to realize the learning status of learners by the e-assessment in e-learning environment. The effectiveness of e-assessment is closely related with the content-richness of the itembank. For increasing the content-richness of itembank the thesis proposed a heterogeneous itembank integrator to integrate several different itembanks which can be with heterogeneous schemata in the same specific knowledge domain. The itembanks in the same specific knowledge domain may be with heterogeneous schemata because the instructors or editors of the textbooks/itembanks are with the various knowledge backgrounds or different philosophies of education. This thesis focuses on designing an effective itembank integrator to integrate the itembanks into an integrated itembank. Essentially, the integration method is based on the vector space model. And the proposed integrating method are improved by the weighted-content strategy and latent semantic indexing. In addition, the structure information of schema is also used to enhance the performance of integration. The experimental result shows the effectiveness of integration. In addition, we can find the influence of different education theories on the itembank integration. With the increasing of content richness, the integrated itembank has more redundancy among items. We also proposed a new method based on genetic algorithm to reduce the effect of redundancy in test sheet assembling. The experimental result shows the availability of the approach. At last, we extended the itembank integrator to the content/item integrator with the ability of integrating the content from Internet for more information richness in the e-assessment feedback. The experimental result also shows that the content/item integrator can provide more useful information in the feedback of e-assessment.