ABSTRACT
ABSTRACT A new method of preliminary swarm intelligence, namely, ant colony algorithm-based classification, is investigated in this chapter. Data were collected from the regions connecting the urban and suburban areas of Fuzhou in China’s Fujian Province, forming a multisource database integrating spectral information, topographical characteristics, and textural information. Classification rules were developed on the basis of differing characteristics in the samples using the ant colony algorithm. In addition, the traditional maximum likelihood method, C4.5 algorithm, and rough set method were also applied for comparison and to check the classification
CONTENTS
16.1 Introduction ................................................................................................ 270 16.2 Study Area and Data ................................................................................. 270
16.2.1 Study Area ...................................................................................... 270 16.2.2 Data Source ..................................................................................... 271
16.3 Method ........................................................................................................ 272 16.4 Comparison and Discussion .................................................................... 274
16.4.1 Classification by Ant Colony Algorithm Based on Different Variables ......................................................................... 274
16.4.2 Select Rules Based on Different Numbers of Variables or Characteristics ................................................................................277