1887

Abstract

This study aims to predict sonic log data using density log, gamma log and porosity. For the purpose Random Forest Regression methods, Support Vector Regression (SVR) algorithm with kernels such as linear, gaussian and polynomial and K Nearest Neighbors algorithm has been applied. SVR linear kernel turns out to be the best algorithm for prediction of sonic data using gamma log, density log and porosity. The training time for SVR linear method is much less than neural networks. In neural networks whereby number nodes in input layer and hidden layers, number of hidden layers required , the functional form in the node responsible for mapping from one node to other node has to be chosen while in SVR linear method only box constraint and margin has to be chosen. One does not have to bother with the order of the model as in case of auto regressive models. Presence of other log data such as resistivity etc could have helped in improving the predictability of sonic log using SVR linear method.

Loading

Article metrics loading...

/content/papers/10.3997/2214-4609.201802090
2018-08-22
2024-03-28
Loading full text...

Full text loading...

http://instance.metastore.ingenta.com/content/papers/10.3997/2214-4609.201802090
Loading
This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error