The Study Of Data Mining In Ubiquitous- Learning Evaluation

The "National Long-term Educational Reform and Development Plan (2010-2020)" states that, "By 2020, China will be a 3A learning society, which means everybody can learn anywhere, anytime, and with any device." It also coincides with the idea of Ubiquitous Learning (U-learning). The construction of a 3A-learning society can help to promote the implementation of the U-learning, which can enable the general public to achieve life-long learning.With the fast development of modern network communications and computer technology, ubiquitous computing, distance education, mobile learning, lifelong learning, and E-learning will become the major components of future education. With the wide development of Ubiquitous Computing Networks, how to properly evaluate different learning models becomes critical in improving the quality of learning at many forms. The evaluation will involve digging for strategy of the assessment data, looking for the key factors which impact the learning assessment. And these will spur a better and faster development of a 3A-learning society. After studying various types of universal open-learning-modes supported by current technologies, the thesis proposes a specific ubiquitous learning model. By connecting the relationship between human subjective consciousness and the objective form of learning resources, the author, in collaboration with experts from San Diego State University, generated the "message design principles for mobile learning". And these principles could be a favorable reference for the follow-up design of U-learning resources. In addition, this article provides a quality assessment method based on E-learning resources platforms and systems, in order to evaluate the effectiveness of the learning resources. The author also conducted experiments on the weighting distribution test to provide examples on U-learning evaluation index-setting, and other experiments on U-learning analysis by data mining technology, providing various reference and cases on process methods for future study on U-learning.The main work of this paper can be divided into four parts as followed:First, summarize the specific context of ubiquitous learning models under the current technology and the different learning forms.Second, after study on the current evaluation system and some mainstream data mining cases, this thesis suggests a set of assessment index on keeping electronic learning materials (E-materials) effective and efficient. This assessment can help to develop sustainable E-materials to meet the potential needs of U-learning. A weighting distribution experiment on this evaluation index set has been put forward, which could be a good reference on weighting coefficients and distribution methods for U-learning evaluation.Third, this thesis provides an effective solution for the sequencing data of subjective evaluation after the study on a variety of data mining algorithms. It also gives an experimental model for future's U-learning data analysis.Finally, this thesis uses neural network-based data mining technology to analyze experiments on the study of learning behavior and learning effectiveness. Factor analysis experiment is used to discover the key factors that impact the learning evaluation results.This thesis provides several solid experiments and can be a good reference on research methods and data mining strategies for the future study on U-leaning.