Cross-validation estimation of regression in seismic interpretation
Abstract
Regression analysis, as the most simple and understandable way to predict parameters, has received a wide coverage, including in geosciences. The error of this method can be estimated by such parameters as correlation coefficient and standard error. Cross-validation (cross-validation) is one of the ways to assess the stability of a model in which part of the input does not participate in the analysis, but it is uses for evaluation. Usually cross-validation is not used in regression analysis, however, using these approaches in aggregate, it is possible to estimate the correlation coefficient and the standard error for each of the implementations, as well as evaluate the contribution of input data to a particular well. It is possible to identify problem points in the initial data at the stage of mapping.
About the Authors
R. A. Sukhanov
Gazpromneft NTC LLC
Russian Federation
Russian Federation
Saint-Petersburg
A. V. Ekimenko
Gazpromneft NTC LLC
Russian Federation
Russian Federation
Saint-Petersburg
References
1. Allen D. M. The relationship between variable selection and data augmentation and a method for prediction // Technometrics. - 1974. – V. 16. – P. 125–127.
2. Geisser S. The predictive sample reuse method with applications // Journal of the American Statistical Association. - 1975. – V. 70. – P. 320–328.
3. Stone M. Cross-validatory choice and assessment of statistical predictions // Journal of the Royal Statistical Society. – Ser. B. - 1974. – V. 36. – P. 111–147.
4. Shao J. Linear model selection by cross-validation // Journal of the American Statistical Association. – 1993. – V. 88(422). – P. 486–494.
5. Arlot S., Celisse A. A survey of cross-validation procedures for model selection // Statistics Surveys. – 2010. - V. 4. - DOI: 10.1214/09-SS054
Review
For citations:
Sukhanov R.A., Ekimenko A.V. Cross-validation estimation of regression in seismic interpretation. PROneft. Professionally about Oil. 2017;(4):15-17. (In Russ.)
Views:
138
Email this article 