However, in practice, it is difficult to identify such fuzzy measures. An automated process is necessary and can be used when sample data is available. Several optimization approaches have been proposed to extract fuzzy measures from sample data; for example, genetic algorithms, gradient descent algorithms, and the Bees algorithm.
In our previous work, instead of using the search space as the primary focus of our research, we developed an algorithm that speculates on the value of the objective function before actually arriving to it. Contrary to previous approaches to extracting fuzzy measures, our algorithm guarantees the solution to be global. However, the improvement is limited by cases in which the value of the objective function could not be further reduced / determined.
In this article we propose an extension to the algorithm, in which local search algorithms, such as hill climbing, are used to provide, at a cheaper cost, insights about the value of the objective function at times when our speculative algorithm fails to progress. The resulting insights are fed back into the speculative algorithm. Our experimental results show that our algorithm improves the performance of previous approaches.