FRULER: Fuzzy Rule Learning through Evolution for Regression

Abstract

In regression problems, the use of TSK fuzzy systems is widely extended due to the precision of the obtained models. Moreover, the use of simple linear TSK models is a good choice in many real problems due to the easy understanding of the relationship between the output and input variables. In this paper we present FRULER, a new genetic fuzzy system for automatically learning accurate and simple linguistic TSK fuzzy rule bases for regression problems. In order to reduce the complexity of the learned models while keeping a high accuracy, the algorithm consists of three stages: instance selection, multi-granularity fuzzy discretization of the input variables, and the evolutionary learning of the rule base that uses the Elastic Net regularization to obtain the consequents of the rules. Each stage was validated using 28 real-world datasets and FRULER was compared with three state of the art genetic fuzzy systems. Experimental results show that FRULER achieves the most accurate and simple models compared even with approximative approaches.

Fuzzy Model of the Operational Potential Consumption Process of a Complex Technical System

“…Modeling of the operational potential consumption process can be treated as a case when adequate mathematical model does not exist and tries of the model construction face the problems arising from not clear enough process description and approximate character of the input data. In such cases the fuzzy modeling implementation can be found in industrial application [22][23][24][25] as well as in theoretical research papers [26][27][28]. Thus, on the basis of the research conducted by the author and the literature analyze it was decided that the MPu model will be constructed by the fuzzy modeling implementation.…”

“…The exemplary form of the obtained inference rule is presented below Eq. (26). In all expressions describing the rule and knowledge bases of the fuzzy models the symbols were used in accordance with tables (Table 2 - Table 5).…”

Study on the use of different quality measures within a multi-objective evolutionary algorithm approach for emerging pattern mining in big data environments

“…There development of evolutionary algorithms for Big Data is a challenge because of its complexity. Nowadays, there are several efforts for the development of evolutionary algorithms in several data mining tasks [30][31][32][33]. To the best of our knowledge, the developed algorithms are EvAEFP-Spark [34] and BD-EFEP [17].…”

Efficiency in BRICS banking under data vagueness: A two-stage fuzzy approach

“…Following an emerging trend in the field (Antonelli, Ducange, Marcelloni, & Segatori, 2016;Rodríguez-Fdez et al, 2016), the present study adopts the second approach, comparing learning methods that are usually classified into different groups (Riza et al, 2015): space partition, clustering, and neural networks. An FRBS can be used just like other regression models and their corresponding packages in R. The principles of the technique are described in Table 2.…”