skip to main content

ANALISA DATA CURAH HUJAN STASIUN KLIMATOLOGI SEMARANG DENGAN MODEL JARINGAN SYARAF TIRUAN

*F M Arif  -  Laboratorium Geofisika, Jurusan Fisika, Universitas Diponegoro, Indonesia
Rahmat Gernowo  -  Laboratorium Geofisika, Jurusan Fisika, Universitas Diponegoro, Indonesia
Agus Setyawan  -  Laboratorium Geofisika, Jurusan Fisika, Universitas Diponegoro, Indonesia
D Febrianty  -  Badan Meteorologi Klimatologi dan Geofisika Jawa Tengah, Indonesia

Citation Format:
Abstract

The major purpose of this research was to applying artificial neural network to predicting rainfall in Semarang climatology station and occurs its accuration. One ofartificial neural network method is back propagation artificial neural network. Withheuristic technique its optimizing to train algorithmic faster and improving net works. Weused rainfall data in 2000-2009 from Semarang climatology station. Artificial neuralnetwork modelling planned in MATLAB R2008b programme. The best model or net viewsfrom correlation level between net’s output, observation data and RMSE point whichproduced by the net. The results shown the best network has 5 neurons in input’s layer, 10in hidden layer and 1 neuron in output layer. Its performance has learning data 66,7%,testing data 33,3%, learning rate 0,7 and momentum 0,4 which has correlated around70,72% to observation data with RMSE point 141,55. The best network will use topredicting rainfalls in 2010, its correlation is 88,43% and its RMSE points is 83,76 tillJuly. Its better than what BMKG has which only reach 84,63% correlation points and87,21 RMSE points.

Keywords:  Artificial neural network, optimizing, correlation, RMSE

Fulltext View|Download

Last update:

No citation recorded.

Last update: 2024-11-20 03:37:34

  1. Recurrent gradient descent adaptive learning rate and momentum neural network for rainfall forecasting

    Kanigoro B.. Proceedings - 2016 International Seminar on Application of Technology for Information and Communication, ISEMANTIC 2016, 2017. doi: 10.1109/ISEMANTIC.2016.7873804