IMPLEMENTASI ADAPTIVE NEURO FUZZY INFERENCE SYSTEM (ANFIS) UNTUK PREDIKSI CURAH HUJAN STUDI KASUS BMKG STAMET DABO

  • Marfuah Marfuah Alumni Program Studi Teknik Informatika
  • Martaleli Bettiza Dosen Fakultas Teknik UMRAH
  • Alena Uperiati Dosen Fakultas Teknik UMRAH
Keywords: Prediction, Rainfall, Normalization, K-means, Adaptive Neuro Fuzzy Inference System (ANFIS

Abstract

The weather conditions in each area in Lingga Regency are monitored by the Meteorology, Climatology and Geophysics Agency (BMKG) of the Dabo Meteorological Station (Stamet) which can change every day, weather is very important and greatly influences the activities of all living things, rainfall is one of the factors of weather which also affects the activities of living things. Therefore it is necessary to know the conditions of rainfall every day. In this case, we need a system that is able to predict rainfall. In this final project, the author uses wind speed, air humidity, temperature, and air pressure data as input, while rainfall will be the target output starting from January 2017 to December 2018. The prediction method used is ANFIS (Adaptive Neuro Fuzzy Inference System) by applying the Fuzzy K-Means method to perform generalized bell fuzzification. In this scientific work ANFIS is used to predict rainfall in the Lingga Regency area. From the results of the tests carried out, it can be seen that in the training data the smallest error is obtained at Learning rate 0.8, namely with a value of 0.024456, while in the new premise parameters in the test data, the smallest error is obtained at a learning rate of 0.9 with a value of 0.01255 using RMSE (Root Mean Squared Error).

Published
2021-10-12
Section
Teknik Informatika