| 摘要 |
An apparatus and method for automatically generating fuzzy rules and/or membership functions for a fuzzy inference system are disclosed in which an input system, constituted by a computer system, is provided having a learning memory for storing learning data for all of input spaces, the input spaces being clustered for each fuzzy inference input variable and an output system, constituted by the computer system, is provided for determining a fuzzy label for each input variable, for dividing support sets of the respective input variables according to the determined fuzzy labels, for determining and storing a weight table of the computer system according to the number of divided support sets, for learning the computer system using the learning data on a basis of the fuzzy labels of the respective input variables, the weight table, and the number of divided support sets so that each representative point of the fuzzy labels of the input variables is input to the learned computer system to automatically derive at least one output value corresponding to the representative points for the input variables, and for providing the fuzzy labels for the output variable to fuzzify the output value to form the membership function of the output system.
|
