IMPLEMENTASI NEURAL FUZZY INFERENCE SYSTEM DAN ALGORITMA PELATIHAN LEVENBERG-MARQUARDT UNTUK PREDIKSI CURAH HUJAN (Studi Kasus: Stasiun Pengamatan Curah Hujan di Putussibau Kalimantan Barat); IMPLEMENTATION OF NEURAL FUZZY INFERENCE SYSTEM AND LEVENBERG-MARQUARDT TRAINING ALGORITHM FOR RAINFALL PREDICTION (Case Study: Rainfall Station in Putussibau West Kalimantan)

Ritha, Nola, Retantyo Wardoyo

2015 | Disertasi | FMIPA

Rainfall prediction can be used for various purposes and the accuracy in predicting is important, especially for specialized fields such as agriculture and industry. In this research, the data of rainfall prediction is used daily rainfall data from the years 2013-2014 at rainfall station in Putussibau, West Kalimantan. Factors affecting precipitation which is used as a parameter. Rainfall prediction using four parameters: mean temperature, average humidity, wind speed and mean sea level pressure. This research to determine how performance Neural Fuzzy Inference System with Levenberg-Marquardt training algorithm for rainfall prediction. Fuzzy logic can be used to resolve linguistic variables that used in rule of rainfall. While artificial neural networks have the ability to adapt and the ability of the learning process (learning), due to recognize patterns of data from the inputs before in predicting rainfall needed training. And Levenberg-Marquardt algorithm is used for training because of the effectiveness and convergence acceleration. The results showed five models NFIS-LM developed using a variety of membership functions, linguistic variables and parameters factor of rainfall as input obtained that the model NFIS-LM with twelve of membership functions and use of three inputs, such as, mean temperature, average humidity, wind speed and mean sea level pressure gives the best results to predict rainfall with values Mean Square Error (MSE) of 0.0262050. When compared with the model NN-Backpropagation, NFIS-LM models showed lower accuracy. It is shown from MSE where the model NN-Backpropagation generate MSE of 0.0167990.

Kata Kunci : Levenberg-Marquardt; Neural Fuzzy Inference System; Rainfall Prediction.