Genetic Rules Induction Fuzzy Inference System for Classification and Regression Application in Energy Industry

Abstract

Genetic fuzzy system encompasses genetic algorithm and fuzzy logic. It divulges the advantage of optimization with ease of understanding for classification and regression of energy performance of buildings, transformer, and harmonic current in energy industry. This paper presents development of a new rules induction algorithm namely genetic rules induction fuzzy inference system for classification and regression (GRIFISCnR) that combines genetic algorithm with fuzzy logic to facilitate efficient design of building, transformer and harmonic current filter in energy industry using Pittsburgh approach. GRIFISCnR possesses the rules induction capability over other algorithms for multi-class classification and regression problems without compromising on interpretability and accuracy. It manages to strike a balance between interpretability and accuracy, and yield better accuracy with lesser number of rules. It is easier to interpret and understand fuzzy rules as compared to numerical numbers.