CAREER: Ubiquitous Distributed Knowledge Discovery from Heterogeneous Data

点击翻译按钮获取中文摘要

The goal of this research project is to develop a new generation of algorithms and systems for mining distributed and heterogeneous data for ubiquitous computing environments. This research involves development of a distributed principal component analysis (PCA), distributed randomized projection techniques, clustering algorithms based on these distributed representation construction techniques, and a distributed decision tree learning technique based on Fourier analysis. This research is integrated with promotion of education in the data mining area through hosting undergraduates, K-12 students and teachers, organizing workshops, maintaining virtual presence, working with under-represented groups in collaboration with the University of Maryland Baltimore County Shriver Center, and the development of research/instructional laboratories. The results of this project will provide a new generation of data mining algorithms that minimize the cost for maintaining the ubiquitous presence. It will also enhance the awareness among the students at different levels about the importance of data mining education.