| Recently, the architecture of intelligent decision support system has been greatly improved when artificial intelligence, including machine learning, data mining, rough set theory, Dempster-Shafer theory and so on, was developed rapidly. At the end of last century and the very beginning of this century, we can acquire more and more information with the popularization of Internet, and plenty of information and convenient communication platform are provided, and more experts can participate in decision activities. Research on intelligent group decision theory and methods nowadays become an active research area.Currently, the key problems of research on intelligent group decision theory and methods mainly include: Firstly, a mass of semi-structured, unstructured, uncertain and correlative information will affect decision. Experts hardly exactly, roundly and fast grasp and analyze all information by the knowledge and experience buried in their mind, and need intelligent methods to help resolving decision problem, and then form a decision opinion; Secondly, there always exists conflicts in lots of experts' opinions. More experts more difficult to solve decision problem sometimes, group decision theory and methods should solve conflicts and investigate the methods of selecting and evaluating the optimal decision opinion.This dissertation investigates and analyzes the research about intelligent decision support system and group decision support system, and then aims at solving decision problem when facing mass of unstructured, correlative information, solving conflicts and studying performance of fusion rules with lots of experts.The main innovations of this dissertation can be described as below:First, To discover and analyze that how correlation among properties of data affects performance of decision tree, a series of comparative algorithms using or eliminating correlation have been designed, and experimental results and its' analysis give the...
|
PDF Full Text Request