Abstract

ABSTRACT Research has been carried out on the analysis of soil acceleration using Artificial Neural Networks (ANN) which aims to analyze and predict soil acceleration in Sumatra. This study uses ground acceleration data recorded through accelerographs found at 3 stations in West Sumatra, namely Padang Panjang Geophysical Station (PAPA), Ketaping Meteorological Station (PATA), and Teluk Bayur Meritim Station (PATU). This data processor uses the backpropagation method. In data processing, there are 2 types of data sharing, namely data sharing with a ratio of 50:50 and a ratio of 80:20. After the training and testing process was carried out, it was found that data sharing with a ratio of 80:20 got better results than sharing data with a ratio of 50:50. Overall, it can be concluded that the ANN in the training process is able to predict ground acceleration quite accurately, but in the testing process, the error value obtained is quite large, so that the ANN is unable to predict the soil acceleration data well. Key words: ground acceleration, artificial neural network, backpropagation

Toni, Widianto (2019) ANALISIS PERCEPATAN TANAH GEMPA DI SUMATERA MENGGUNAKAN JARINGAN SYARAF TIRUAN METODE BACKPROPAGATION. Diploma thesis, Universitas Andalas.

(Link)