Details
| Original language | English |
|---|---|
| Pages (from-to) | 71-94 |
| Number of pages | 24 |
| Journal | Structural Engineering and Mechanics |
| Volume | 32 |
| Issue number | 1 |
| Publication status | Published - 10 May 2009 |
| Externally published | Yes |
Abstract
In this paper a procedure for Monte Carlo simulation of univariate stationary stochastic processes with the aid of neural networks is presented. Neural networks operate model-free and, thus, circumvent the need of specifying a priori statistical properties of the process, as needed traditionally. This is particularly advantageous when only limited data are available. A neural network can capture the "pattern" of a short observed time series. Afterwards, it can directly generate stochastic process realizations which capture the properties of the underlying data. In the present study a simple feedforward network with focused time-memory is utilized. The proposed procedure is demonstrated by examples of Monte Carlo simulation, by synthesis of future values of an initially short single process record.
Keywords
- Monte Carlo simulation, Neural networks, Stochastic processes
ASJC Scopus subject areas
- Engineering(all)
- Civil and Structural Engineering
- Engineering(all)
- Building and Construction
- Engineering(all)
- Mechanics of Materials
- Engineering(all)
- Mechanical Engineering
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: Structural Engineering and Mechanics, Vol. 32, No. 1, 10.05.2009, p. 71-94.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - A neural network approach for simulating stationary stochastic processes
AU - Beer, Michael
AU - Spanos, Pol D.
PY - 2009/5/10
Y1 - 2009/5/10
N2 - In this paper a procedure for Monte Carlo simulation of univariate stationary stochastic processes with the aid of neural networks is presented. Neural networks operate model-free and, thus, circumvent the need of specifying a priori statistical properties of the process, as needed traditionally. This is particularly advantageous when only limited data are available. A neural network can capture the "pattern" of a short observed time series. Afterwards, it can directly generate stochastic process realizations which capture the properties of the underlying data. In the present study a simple feedforward network with focused time-memory is utilized. The proposed procedure is demonstrated by examples of Monte Carlo simulation, by synthesis of future values of an initially short single process record.
AB - In this paper a procedure for Monte Carlo simulation of univariate stationary stochastic processes with the aid of neural networks is presented. Neural networks operate model-free and, thus, circumvent the need of specifying a priori statistical properties of the process, as needed traditionally. This is particularly advantageous when only limited data are available. A neural network can capture the "pattern" of a short observed time series. Afterwards, it can directly generate stochastic process realizations which capture the properties of the underlying data. In the present study a simple feedforward network with focused time-memory is utilized. The proposed procedure is demonstrated by examples of Monte Carlo simulation, by synthesis of future values of an initially short single process record.
KW - Monte Carlo simulation
KW - Neural networks
KW - Stochastic processes
UR - http://www.scopus.com/inward/record.url?scp=67449163735&partnerID=8YFLogxK
U2 - 10.12989/sem.2009.32.1.071
DO - 10.12989/sem.2009.32.1.071
M3 - Article
AN - SCOPUS:67449163735
VL - 32
SP - 71
EP - 94
JO - Structural Engineering and Mechanics
JF - Structural Engineering and Mechanics
SN - 1225-4568
IS - 1
ER -