Spatial Data Mining: Theory and Application

·        This book is an updated version of a
well-received book previously published in Chinese by Science Press of China
(the first edition in 2006 and the second in 2013). It offers a systematic and
practical overview of spatial data mining, which combines computer science and
geo-spatial information science, allowing each field to profit from the
knowledge and techniques of the other. To address the spatiotemporal
specialties of spatial data, the authors introduce the key concepts and
algorithms of the data field, cloud model, mining view, and Deren Li methods.
The data field method captures the interactions between spatial objects by
diffusing the data contribution from a universe of samples to a universe of
population, thereby bridging the gap between the data model and the recognition
model. The cloud model is a qualitative method that utilizes quantitative
numerical characters to bridge the gap between pure data and linguistic
concepts. The mining view method discriminates the different requirements by
using scale, hierarchy, and granularity in order to uncover the anisotropy of
spatial data mining. The Deren Li method performs data preprocessing to prepare
it for further knowledge discovery by selecting a weight for iteration in order
to clean the observed spatial data as much as possible. In addition to the
essential algorithms and techniques, the book provides application examples of
spatial data mining in geographic information science and remote sensing. The
practical projects include spatiotemporal video data mining for protecting
public security, serial image mining on nighttime lights for assessing the
severity of the Syrian Crisis, and the applications in the government project
‘the Belt and Road Initiatives’.