Rainfall prediction by using ANFIS times series technique in South Tangerang, Indonesia
Wayan Suparta, Azizan Abu Samah
Abstract
Excessive rainfall is one of the triggers for the flooding phenomenon, especially in the tropics with flat or concave areas. Some critical points in the South Tangerang region, which are currently one of the most rapidly developing cities, cannot be ignored from the flooding problem. Floods cause disturbing human activities, loss of life and property, and in turn affect the economic stretch in an area. This paper aimed to predict rainfall by exploring the application of artificial intelligence techniques such as ANFIS (Adaptive NeuroFuzzy Inference System). The proposed technique combines neural network learning abilities with transparent linguistic representations of fuzzy systems. The ANFIS model with various input structures and membership functions was built, trained, and tested to evaluate the capability of a model. Analyses of six-year rainfall data on a monthly basis in South Tangerang City, Banten found that rainfall prediction based on ANFIS time series is promising where 80% of data testing is well predicted.