Extracting linguistic knowledge and its use as decision support in a virtual futures market

In this paper, we try to extract useful linguistic knowledge from an autonomous agent in a virtual futures market. An adaptive fuzzy rule-based system is used for an autonomous agent in order to make a decision of the trade action such as buy the futures index or sell. The adaptive fuzzy rule-based system adjusts the weights of fuzzy if-then rules such that the trade performance of the autonomous agent can be gradually improved. We select a small number of fuzzy if-then rules from the adjusted fuzzy rule-based systems to form a knowledge base for the futures market. In our experiments, we examine the effectiveness of the knowledge base as a decision support tool for human beings in a futures market by presenting the knowledge base to the human beings. The experimental results along with the result of a statistical test show that human beings can perform better decision making in the futures market with the help of the knowledge base extracted from the adaptive fuzzy rule-based system.