Enhancing the performance of smart electrical grids using data mining and fuzzy inference engine

Abstract This paper is about enhancing the smart grid by proposing a new hybrid feature-selection method called feature selection-based ranking (FSBR). In general, selection to exclude non-promising features out from collected data at Fog. could be achieved using filter methods, wrapper or hybrid. Our proposed consists of two phases: and phases. phase, whole go through different techniques (i.e., relative weight ranking, effectiveness information gain ranking) The results these ranks are sent fuzzy inference engine generate final ranks. being selected based on passed three classifiers Naive Bayes, Support Vector Machine, neural network) select best set performance classifiers. process can enhance reducing amount cloud, decreasing computation time, complexity. Thus, FSBR methodology enables user load forecasting (ULF) take fast decision, reaction in short-term forecasting, provide high prediction accuracy. authors explain suggested approach via numerical examples. Two datasets used applied experiments. first dataset reported that was compared with six other represented accuracy 91%. second set, generalization 90% fourteen methods.