Secondary language: |
English |
Secondary title: |
ǂThe ǂprediction of landslide movements with RBF artificial neural networks |
Secondary abstract: |
The thesis deals with the problem of the prediction of landslide movements with artificial
neural networks (ANN). Landslide movements are dependent on many parameters (rainfall,
topography, geology, material, ...), what makes them hard to predict. The links between
movements, rainfall, temperature and surface inkilnation are too complex to define them with
mathematical equations. That is why we use artificial neural networks. In the beginning we
describe the Macesnik landslide, where the observations took place. We continue with the
description of previous observations on the Macesnik landslide and we analyze the influence
of rainfall on movements of the landslide. Next, we describe artificial neural networks,
distribution of the artificial neural networks, the criteria of distributing artificial neural
networks and present a detail description of the radial basis function (RBF) neural networks.
In the experimental part we have presented the use of artificial neural networks for the
prediction of landslide movements. For input and output data we have used the measurements
of rainfall and measured movements of the landslide, respectively. Our goal is to find a
artificial neural network, which could be used for landslide movement prediction in the future.
For the calculation we used two different sotwares – Matlab R2007b and Neurosolutions. At
the end conclusions are given, which include a short summary of all major findings. |
Secondary keywords: |
graduation thesis;geodesy;Macesnik's;landslides;artificial neural network;RBF;geodetic measurements;movements;rainfall;analysis; |
File type: |
application/pdf |
Type (COBISS): |
Undergraduate thesis |
Thesis comment: |
Univ. v Ljubljani, Fak. za gradbeništvo in geodezijo |
Pages: |
XII, 80 str., [4] pril. |
Type (ePrints): |
thesis |
Title (ePrints): |
The prediction of landslide movements with RBF artificial neural networks |
Keywords (ePrints): |
Macesnikov plaz;umetna nevronska mreža;RBF;geodetske
meritve;premiki;padavine;analiza |
Keywords (ePrints, secondary language): |
Macesnik’s landslide;artificial neural network;RBF;geodetic
measurements;movements;rainfall;analysis |
Abstract (ePrints): |
Diplomska naloga predstavlja možnost uporabe umetnih nevronskih mrež pri napovedovanju
premikov pri plazenju tal. Premiki plazov so odvisni od mnogih dejavnikov (padavine,
topografija, geologija, materialne lastnosti, …), zato jih je težko napovedati. Povezave med
premiki, padavinami, temperaturami in inklinacijo terena so preveč kompleksne, da bi jih
opisali z matematičnimi formulami. Zato jih poskusimo opisati z umetnimi nevronskimi
mrežami. V nalogi najprej predstavimo Macesnikov plaz, kjer so bile izvedene meritve. V
nadaljevanju opišemo dosedanja opazovanja Macesnikovega plazu in analiziramo vpliv
padavin na plazenje zemljin. Sledi predstavitev umetnih nevronskih mrež, razdelitev umetnih
nevronskih mrež, kriteriji razdelitve umetnih nevronskih mrež in podrobna razložitev
radialnih bazičnih umetnih nevronskih mrež. V eksperimentalnem delu je prikazana uporaba
radialnih bazičnih umetnih nevronskih mrež pri napovedovanju plazenja zemljin. Za vhodne
podatke smo uporabili meritve padavin, za izhodne pa premike plazenja zemljine. Cilj je
dobiti naučeno umetno nevronsko mrežo, ki bi jo lahko uporabili za napovedovanje premikov
v praksi. Za izračun uporabimo dve različni programski opremi – Matlab R2007b in
Neurosolutions, katerih rezultate primerjamo med seboj. Na koncu podamo zaključek, ki
vsebuje kratek povzetek vseh glavnih ugotovitev. |
Abstract (ePrints, secondary language): |
The thesis deals with the problem of the prediction of landslide movements with artificial
neural networks (ANN). Landslide movements are dependent on many parameters (rainfall,
topography, geology, material, ...), what makes them hard to predict. The links between
movements, rainfall, temperature and surface inkilnation are too complex to define them with
mathematical equations. That is why we use artificial neural networks. In the beginning we
describe the Macesnik landslide, where the observations took place. We continue with the
description of previous observations on the Macesnik landslide and we analyze the influence
of rainfall on movements of the landslide. Next, we describe artificial neural networks,
distribution of the artificial neural networks, the criteria of distributing artificial neural
networks and present a detail description of the radial basis function (RBF) neural networks.
In the experimental part we have presented the use of artificial neural networks for the
prediction of landslide movements. For input and output data we have used the measurements
of rainfall and measured movements of the landslide, respectively. Our goal is to find a
artificial neural network, which could be used for landslide movement prediction in the future.
For the calculation we used two different sotwares – Matlab R2007b and Neurosolutions. At
the end conclusions are given, which include a short summary of all major findings. |
Keywords (ePrints, secondary language): |
Macesnik’s landslide;artificial neural network;RBF;geodetic
measurements;movements;rainfall;analysis |
ID: |
8312117 |