Journal of Data Analysis
Volume 1, Number 2, December 2018

Penerapan Neural Network Backpropagation dengan Transformasi Wavelet Morlet Data Rata-Rata Pasang Surut Air Laut Di Pantai Ulee Lheue

Novira Iswani (Jurusan Statistika, FMIPA, Universitas Syiah Kuala)
Ichsan Setiawan (Jurusan Statistika, FMIPA, Universitas Syiah Kuala)
Miftahuddin Miftahuddin (Jurusan Statistika, FMIPA, Universitas Syiah Kuala)



Article Info

Publish Date
23 Dec 2018

Abstract

Pasang surut berpengaruh terhadap pengoptimalan dan pemanfaatan potensi laut dan segala aktifitas yang akan dilakukan di laut, terutama aktifitas di tepi pantai. Sehingga diperlukan pendeteksian fenomena alam yang mungkin terjadi terutama di daerah yang rawan bencana seperti Aceh. Penelitian ini menggunakan metode Neural network backpropagation yang difokuskan pada pemodelan kondisi pasang surut. Pemodelan dilakukan dengan menerapkan teori markov chain dan untuk memperoleh model terbaik data di transformasi menggunakan transformasi wavelet morlet. Penerapan neural network backpropagation dalam menggunakan data pasang surut di pantai Ulee Lheue, Banda Aceh periode tahun 2013-2017. Terdapat 5 variabel yang digunakan dalam penelitian, yaitu pasang surut yang terjadi di pagi, siang, sore, malam dan dini hari. Tujuan dari penelitian adalah untuk memperoleh model terbaik dari pasang surut air laut di pantai Ulee Lheue menggunakan neural network backpropagation. Hasil yang diperoleh menunjukkan bahwa model jaringan dengan input 1, hidden 2 dan output 1 atau model jaringan 1–2–1  merupakan model terbaik neural network backpropagation.Tides affect the optimization and utilization of the potential of the sea and all activities that will be carried out at sea, especially activities on the beach. So that it is possible to detect natural phenomena that might occur especially in disaster-prone areas such as in Acehness. This research uses the neural network backpropagation method which is focused on modeling. Modeling is done by applying the Markov Chain theory and to obtain the best model the data is transformed using the morlet wavelet transform. Application of neural network backpropagation in uses tidal data on the coast of Ulee Lheue, Banda Aceh for the 2013-2017 period. There are five variables used in research are tide that occur in the morning, afternoon, evening, night and early morning. The purpose of research is to obtain the best model from tides of Ulee Lheue use neural network backpropagation. The results obtained show that the network model with input 1, hidden 2 and output 1 or 1–2–1 network model is the best model of backpropagation neural network.

Copyrights © 2018






Journal Info

Abbrev

JDA

Publisher

Subject

Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Mathematics

Description

Journal of Data Analysis (JDA) is a journal which has scope in Actuary, Algebra, Applied Mathematics, Applied Statistics, Big Data, Biostatistics, Business and Industrial Statistics, Calculus, Categorical Data Analysis, Computer Science, Data Mining, Data Science, Classification, Econometrics, ...