An Improved Ant Algorithm For Fuzzy Data Mining

In the past, two mining algorithms were proposed to find suitable membership functions for fuzzy association rules based on the ant colony systems. In the two approaches, the coding of the possible solutions is by binary strings, which form a discrete solution space. The paper extends the original approaches to continuous search space, and a fuzzy mining algorithm based on the improved ant approach is proposed. The improved ant approach doesn't have fixed paths and nodes and produces some paths in a dynamic way according to the distribution functions of pheromones. The experimental results show that the mining process based on the improved ant approach gets better results than that based on the previous two algorithms.