Comparing Stochastic Simulation Algorithms by Measures of Local Accuracy and Precision
Abstract/Contents
- Abstract
- There are a large number of geostatistical approaches to model the spatial variability of petrophysical properties. For a given problem, different approaches may be appropriate depending on the goals of the study. An important question occurs: among many geostatistical methods, which one is the "best" for a particular problem at hand?The conventional approach to test an estimation algorithm is to consider cross validation and look at summary statistics such as mean absolute error, mean squared error, or the correlation coefficient between the estimated values and the true values. The cross plot of the true values versus the estimated values reveals the "goodness" of the estimation technique.In the context of stochastic simulation, these measures are on longer relevant since we have a distribution of simulated values. So, an approach that can measure the "goodness" of the multiple realizations is needed. This study presents a new approach based on the "leave-one-out" cross validation method. The first step is to build multiple realizations at the location where a true value is temporarily removed. Secondly, a local conditional cumulative distribution function (ccdf) model is built and four measures, i.e., accuracy (A), precision (P ), goodness(G) and uncertainty(U) are specifically designed to establish the "goodness" of this local ccdf model. The advantages and weaknesses of these new measures are demonstrated by comparing them to the conventional cross validation tools. Furthermore, this study describes the methodology to understand strengths and shortcomings of different geostatistical simulation approaches and to assess them in relative accuracy, precision, and uncertainty.
Description
| Type of resource | text |
|---|---|
| Date created | May 1997 |
Creators/Contributors
| Author | Mo, Yexiang |
|---|---|
| Primary advisor | Deutsch, Clayton V. |
| Degree granting institution | Stanford University, Department of Petroleum Engineering |
Subjects
| Subject | School of Earth Energy & Environmental Sciences |
|---|---|
| Genre | Thesis |
Bibliographic information
Access conditions
- Use and reproduction
- User agrees that, where applicable, content will not be used to identify or to otherwise infringe the privacy or confidentiality rights of individuals. Content distributed via the Stanford Digital Repository may be subject to additional license and use restrictions applied by the depositor.
Preferred citation
- Preferred Citation
- Mo, Yexiang. (1997). Comparing Stochastic Simulation Algorithms by Measures of Local Accuracy and Precision. Stanford Digital Repository. Available at: https://purl.stanford.edu/sv516tq4696
Collection
Master's Theses, Doerr School of Sustainability
View other items in this collection in SearchWorksContact information
- Contact
- brannerlibrary@stanford.edu
Also listed in
Loading usage metrics...