On Applying Random Oracles to Fuzzy Rule-Based Classifier Ensembles for High Complexity Datasets

Fuzzy rule-based systems suffer from the so-called curse of dimensionality when applied to high complexity datasets, which consist of a large number of variables and/or examples. Fuzzy rule-based classifier ensembles have shown to be a good approach to deal with this kind of problems. In this contribution, we would like to take one step forward and extend this approach with two variants of random oracles with the aim that this classical method induces more diversity and in this way improves the performance of the system. We will conduct exhaustive experiments considering 29 UCI and KEEL datasets with high complexity (considering both a number of attributes as well as a number of examples). The results obtained are promising and show that random oracles fuzzy rule-based ensembles can be competitive with random oracles ensembles using state-of-the-art base classifiers in terms of accuracy, when dealing with high complexity datasets.