PROCESS OPERATION SUPPORTING SYSTEM AND OPERATION SUPPORTING METHOD

摘要

PURPOSE:To use even buried rules by analyzing the coupling state in a neural net after learning and calculating the strength of causal relations between input and output factors so as to acquire only rules important for process operation. CONSTITUTION:An operation supporting rule extracting means 180 is provided with an intra-neural net coupling strength calculating means which analyzes the coupling state in the neural net obtained by a learning means 160 to calculate the strength of causal relations between input and output factors and an input and output factor selecting means which select input and output factors to be converted to rules based on obtained strength of coupling. The learning means (neural net module) 160 is provided with the neural net of a hierarchical structure consisting of an input layer, an intermediate layer, and an output layer. In this case, the relative strength of causal relations between input and output factors is calculated to select factors having significant causal relations when the coupling state in the neural net after learning is analyzed to acquire fuzzy rule or crisp rules.