ABSTRACT
This chapter, based on cloud model and Gaussian transformation (GT), implements a Gaussian cloud transformation (GCT), which will convert the data distribution in the problem domain into a number of concepts with different granularities. GCT can be used to address the challenges of generating, selecting, and optimizing a concept's number, granularity, and level in variable granular computing, and to provide a new method for data clustering of big data. Scale, level, and granularity are fundamental terminologies in granular computing that serve as the basis for the research of concept. From different concept hierarchies to analyzing and processing data in the domain space, especially for big data, it helps to understand the amount of information in different granularities. The difficult problem in granular computing research is how to simulate the human adaptability for realizing a concept's classification and clustering. The chapter proposes an adaptive Gaussian cloud transformation (A-GCT) algorithm, to automatically find the appropriate number of concepts with little overlap.
