Adaptive Neuro-fuzzy Inference System (ANFIS) for a multi-campus institution energy consumption forecast in South Africa

University campus as a service industry consumes considerable amount of energy, most especially those with multiple campuses. This study develops four ANFIS models for four campuses of an institution located in South Africa using five climatic data as inputs against energy consumption. The clustering method is fundamental to the feasibility and tractability of ANFIS model. The study explores two clustering techniques- fuzzy c-means (FCM) and grid partition (GP) for data clustering. Their forecast accuracy and computational efficiency were compared. FCM gave a better-forecast accuracy and higher computational efficiency in terms of the CPU time compared to the GP technique. The FCM clustering technique was recommended for use in ANFIS model, where similar time series data is used, due to its accuracy and lesser computational time.