Li, Zhe

Postdoctoral Fellow

Zhe Li has a wide range of research interests in GIS, spatial analysis, remote sensing, machine learning, land use/cover change modeling, physical geography and landscape ecology. He holds a B.S. in Geography from Northeast Normal University, China, a M.S. in Ecology from the Chinese Academy of Sciences and a M.A. in Geography from Clark University. He earned his Ph.D. in Geography in 2007 from the Graduate School of Geography, Clark University. His doctoral research work focused on the development of non-parametric algorithms for neural network models in the use of remotely sensed imagery soft classification and spatial uncertainty analysis. His dissertation proposed Commitment and Typicality measures for both the Kohonen’s Self-Organizing Map (SOM) and Fuzzy ARTMAP neural networks, and explored the relationship between the Commitment measures and Bayesian posterior probabilities, and between the Typicality measures and Mahalanobis method. His research bridges traditional statistical approaches and machine learning techniques through developing these soft classification algorithms. Before he joined SDSU, he worked as a part-time research assistant and a programmer for Clark Labs of Clark Univeristy, where he was responsible for developing and designing NEURALNET modules for IDRISI 14.0 Kilimanjaro and IDRISI 15.0 Andes. Dr. Li integrates GIS and machine learning techniques in understanding human-environment relationship and environmental modeling (e.g., land cover change modeling, species habitat distribution modeling and impacts of climate change on decision making, etc). He is also interested in exploratary spatial data analysis and spatial statistics. He developed a spatial point pattern analysis program (SPPA) for his former GIS class in the COPACE of Clark University. Interested people are welcome to contact him to get it for free.

Dr. Li worked as a post-doctoral research associate with Dr. Matthew Hansen on the NASA funded project and conducting research on global-scale land cover change dynamics by mass-processing remotely sensed data sets and using machine learning land cover characterization methods.

Last modified: 
Jul 28, 2014