Application of anfis technique for prediction of rainfall in the south tangerang city

Abstract

Flood phenomenon is an annual scourge that often occurs at several critical points in the territory of Indonesia, one of which is in the South Tangerang City, Banten. This paper aims to predict rainfall as an early detection model by exploring the application of artificial intelligence techniques such as the Adaptive Neuro-Fuzzy Inference System (ANFIS). The proposed technique was built with various input structures and membership functions, then trained and tested to evaluate its capabilities. Daily meteorological data such as relative humidity, temperature, and partial pressure of water vapour are chosen as intermediate parameters to predict rainfall. While rainfall measurement data is used as a comparison or validation. By analysing the three scenarios that have been carried out, it was found that rainfall can be estimated well. The prediction results with ANFIS Fuzzy-C-Means (FCM) algorithm are promising to detect floods that can later be integrated with more complex systems, especially types of floods at various points in the City of South Tangerang and surrounding areas.