Spatial Data Mining and Knowledge Discovery

Author: Di Kaichang
Publisher:
Publish Date: 2003-03-01
Features: This book primarily uses Cloud Theory, Rough Set Theory, and Inductive Learning Methods as its main research methods, with intelligent GIS analysis and automatic interpretation of remote sensing images as its primary application objectives. It systematically investigates the theories and methods of spatial data mining and knowledge discovery. Data Mining and Knowledge Discovery (DMKD) is a cutting-edge technology in the field of information technology that emerged in the 1990s. It was developed in the context where the rapid growth of data and databases far exceeded human capabilities in data processing and understanding. It is also a result of the integration and convergence of database technology, artificial intelligence technology, statistical technology, visualization technology, and other disciplines. Its purpose is to extract implicit, previously unknown but potentially useful information and knowledge from large volumes of incomplete, noisy, and fuzzy data, providing intelligent and automated means for data and database processing and understanding. Spatial Data Mining and Knowledge Discovery (SDMKD, for short) is the process of extracting implicit spatial and non-spatial patterns and general features of interest from spatial databases. Compared to the information retrieved and queried from spatial database management systems, the knowledge discovered by spatial data mining techniques is implicit, refined, high-level, and of greater value. On one hand, it can enhance the intelligence level of spatial data analysis and application; on the other hand, it can be used to support the automatic interpretation and classification of remote sensing images, promoting the intelligent integration of remote sensing and GIS. SDMKD is not only a key technology in the "3S" integration but also an important component of the "Digital Earth" technology system. This book primarily uses Cloud Theory, Rough Set Theory, and Inductive Learning Methods as its main research methods, with GIS intelligent analysis and remote sensing image automatic interpretation as its primary application objectives, systematically investigating the theories and methods of spatial data mining and knowledge discovery.
Chapter 1 introduces the research status of data mining and knowledge discovery, as well as the demand for knowledge discovery in the fields of remote sensing and GIS. Chapter 2 proposes a theoretical and technical framework for spatial data mining and knowledge discovery, serving as the methodological guidance for the entire research. Chapter 3 systematically studies and develops the cloud model, virtual cloud, cloud transformation, and uncertainty reasoning in Cloud Theory, proposing a series of new models and algorithms. These provide effective solutions for fundamental issues in SDMKD, such as concept and knowledge representation, quantitative-qualitative transformation, generating concepts and concept hierarchies from data, and attribute generalization. Chapter 4 applies Cloud Theory to spatial concept representation, extracting association rules from spatial databases, and conducting uncertain spatial data queries. Chapter 5 introduces Rough Set Theory into the field of GIS, summarizing and organizing a set of methods for attribute analysis and knowledge discovery using Rough Set Theory in GIS. Chapter 6 studies the implementation of inductive learning methods in spatial databases and their applications in remote sensing image classification and intelligent GIS analysis. It proposes a set of inductive learning-based remote sensing image classification techniques and processes, as well as a flexible and general exploratory inductive learning method. Chapter 7 proposes a spatial data clustering algorithm based on mathematical morphology. Chapter 8 provides a summary and outlook for the entire book.
In 1995, my supervisors, Academicians Li Deren and Li Deyi, had a visionary outlook and guided me into the cutting-edge and hot field of Data Mining and Knowledge Discovery. During the research process of my thesis, they devoted a great deal of effort. Professor Li Deren, despite his busy schedule, spent a significant amount of time providing specific guidance on methodological research and application experiments. Professor Li Deyi regularly offered guidance on Cloud Theory and its application in data mining, and engaged in in-depth discussions with me on many detailed issues. On the occasion of the publication of this book, I extend my heartfelt gratitude and sincere respect to them!
During the past few years of study and research, many teachers, leaders, classmates, and colleagues from the School of Information Engineering, Wuhan University of Surveying and Mapping (now Wuhan University of Technology), the State Key Laboratory of Remote Sensing Information Engineering, the Graduate School, and the author's affiliated unit, the Center for Aerial Geophysical and Remote Sensing, have provided much care, guidance, and assistance. I express my sincere gratitude to them here!
This research was supported by the project fund of the State Key Laboratory of Remote Sensing Information Engineering (Project No.: WKL(97)0302) and the Excellent National Key Laboratory Research Project of the National Natural Science Foundation of China (Project No.: 40023004).
Although this book has made some progress, it is undoubtedly preliminary for the emerging field of spatial data mining and knowledge discovery. Due to the limitations of my abilities, there may be many omissions and shortcomings in this book. I sincerely welcome criticism and corrections.