Effect of co-operative fuzzy c-means clustering on estimates of three parameters AVA inversion |
| |
| Authors: | Rajesh R Nair Suresh Ch Kandpal |
| |
| Institution: | (1) School of Biological Sciences, University of Canterbury, Christchurch, New Zealand |
| |
| Abstract: | We determine the degree of variation of model fitness, to a true model based on amplitude variation with angle (AVA) methodology
for a synthetic gas hydrate model, using co-operative fuzzy c-means clustering, constrained to a rock physics model. When
a homogeneous starting model is used, with only traditional least squares optimization scheme for inversion, the variance
of the parameters is found to be comparatively high. In this co-operative methodology, the output from the least squares inversion
is fed as an input to the fuzzy scheme. Tests with co-operative inversion using fuzzy c-means with damped least squares technique
and constraints derived from empirical relationship based on rock properties model show improved stability, model fitness
and variance for all the three parameters in comparison with the standard inversion alone. |
| |
| Keywords: | |
| 本文献已被 SpringerLink 等数据库收录！ |
|
