Spatial Data Analysis And Application Based On Bayesian Methods

Spatial data mining is data mining and knowledge discovery in a new field of study, the data classification is an important branch of the spatial data mining. Bayesian learning theory combines prior information with sample knowledge, but also has a very firm theoretical support, and has been widely studied and applied.In this dissertation, we analyze the defects and problems of weighted naive Bayesian classification algorithm. On this basis, build weighted naive Bayesian classification model using partial least squares regression method based on rough set.Main contents are as follows:In chapter one, we set out the background and significance of the research of this subject, the main content of the naive Bayesian classifier model status and research.In chapter two, we introduce the theoretical basis of the spatial data mining, the elaborated architecture and the basic process of spatial data mining, as well as various classification method in the spatial data mining.In chapter three, we assume the theoretical knowledge of the Bayesian approach, and introduce classification principles and classification process of naive Bayesian classifier model, weighted naive Bayesian classifier model, and analyze the advantages and disadvantages of various models.In chapter four, this chapter presents reduction method based on attribute importance in the rough set, propose partial least-squares regression (PLS) algorithm based on the attribute reduction, determine the weight of the weighted Bayesian classifier algorithm, build the weighted Bayesian classification algorithm base on the attribute reduction PLS. Experimental results show that the proposed new algorithm has higher classification accuracy.In chapter five, the new algorithm is applied to practical problems, compared with other classification algorithms by simulation. In chapter six, the main works of this paper are summarized, and further research questions are indicated.