Cascaded-ANFIS and its Successful Real-World Applications

Abstract

The cascaded adaptive neuro-fuzzy inference system (ANFIS) is widely employed for modeling and controlling nonlinear systems. It combines human experience and knowledge with neural networks’ learning capability and fuzzy systems’ pattern extraction ability. This integration enables the development of effective models across diverse application domains. This chapter introduces the Cascaded-ANFIS algorithm and its case studies. One example of a case study that uses Cascaded-ANFIS is the modeling of the relationship between rainfall and runoff. This relationship is inherently complex and nonlinear, influenced by watershed topography, soil infiltration characteristics, and rainfall patterns. Accurately capturing this relationship is crucial for flood forecasting and water resources management applications. Rainfall data is the primary input variable when employing Cascaded-ANFIS to model the relationship between rainfall and runoff. This encompasses rainfall data with both temporal and spatial resolutions. Runoff data is collected by observing groundwater levels, river water levels, and geographical features of the watershed.