An exchange data analysis support system by intuitive reasonings based on neural network
Kyuichiro TANI, K. Kamei · 2004
The purpose of our work is to design an exchange analysis support system by intuitive reasoning using neural networks. The intuitive reasoning is totally different from logical reasoning used in AI and is defined by subjective information such as high-low, good-bad, etc. A soft computing aided management information analysis system using fuzzy theory, neural networks and intuitive reasoning is a new challenge to management science specialists. In this paper, we propose a new analysis support system based on intuitive reasoning. Firstly, we describe intuitive reasoning. Secondly, intuitive reasoning is applied to exchange rate analysis. Thirdly, the results of intuitive reasoning, that is, subjective exchange rate evaluations and their uncertainty factors are projected on the optimistic-pessimistic axis. Fuzzy membership functions are used for the projection. The total subjective evaluations of the exchange rate are given by a contour line map. Finally, some evaluation results based on actual data of domestic exchange are shown and their analyses are given for system users.