APPLICATION OF GENERAL REGRESSION NEURAL NETWORK IN TSUNAMI DATABASE SYSTEM AT SOUTHERN JAVA
Development of tsunami data base system is one of disaster mitigation efforts, in which there is a possibility to predict maximum tsunami heights and tsunami arrival times immediately along coastline after an earthquake occured. The first step in developing tsunami data base system is to generate hy...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/16148 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Development of tsunami data base system is one of disaster mitigation efforts, in which there is a possibility to predict maximum tsunami heights and tsunami arrival times immediately along coastline after an earthquake occured. The first step in developing tsunami data base system is to generate hypothetic earthquakes base on tsunami historical data in study area. Results from tsunami propagation model using tsunami sources using numerical computation from hypothetic earthquakes as maximum tsunami heights and tsunami arrival times along coastal area were stored into database. The developed database is then used as input and target for neural network learning procces. The type of neural network used in this tsunami data base system is General Regression Neural Network (GRNN) that firstly developed by Specht in 1991.<p>Application of tsunami data base system using GRNN on ideal cases shows quite satisfactory, in which relative error from numerical results based on Root Mean Square Error calculation is less than 6%. Predicting results of GRNN on the real cases indicate smaller error with only less than 3%, this occurred because in the real cases there are more learning data than ideal cases. Computation speed for both cases are almost the same, which is less than 2 second without considering time for data input processes. <br />
|
|---|