A First Approach in Evolutionary Fuzzy Systems based on the Lateral Tuning of the Linguistic Labels for Big Data Classification

The treatment and processing of Big Data problems imply an essential advantage for researchers and corporations. This is due to the huge quantity of knowledge that is hidden within the vast amount of information that is available nowadays. In order to be able to address with such volume of information in an efficient way, the scalability for Big Data applications is achieved by means of the MapReduce programming model. It is designed to divide the data into several chunks or groups that are processed in parallel, and whose result is "assembled" to provide a single solution. Focusing on classification tasks, Fuzzy Rule Based Classification Systems have shown interesting results with a MapReduce approach for Big Data. It is well known that the behaviour of these type of systems can be further improved in synergy with Evolutionary Algorithms, leading to Evolutionary Fuzzy Systems. However, to be best of our knowledge there are no developments in this field yet. In this work, we propose a first Evolutionary Fuzzy System for Big Data problems. It consists of an initial Knowledge Based build by means of the Chi-FRBCS-BigData algorithm, followed by a genetic tuning of the Data Base by means of the 2-tuples representation. This way, the fuzzy labels will be better contextualized within every subset of the problem, and the coverage of the Rule Base will be enhanced. Then, the Knowledge Bases from each Map process are joined to build a ensemble classifier. Experimental results show the improvement achieved by this model with respect to the standard Chi-FRBCS-BigData approach, and opens the way for promising future work on the topic.