Page 255 - Machine Learning for Subsurface Characterization

Chapter 8

Comparative study of shallow

and deep machine learning
models for synthesizing in situ

NMR T2 distributions

Siddharth Misra* and Hao Li †
*
Harold Vance Department of Petroleum Engineering, Texas A&M University, College Station,
†
TX, United States, The University of Oklahoma, Norman, OK, United States
Chapter outline
1 Introduction 220 4.2 Generative adversarial
2 Dataset 222 network assisted neural
3 Shallow-learning models 222 network 230
3.1 Ordinary least squares 222 4.3 Variational autoencoder
3.2 Least absolute shrinkage with convolutional layer
and selection operator 224 assisted neural network 233
3.3 Elasticnet 224 4.4 Encoder-decoder long
3.4 Support vector regressor 225 short-term memory network 233
3.5 k-Nearest neighbor regressor 225 4.5 Comparisons of the accuracy
3.6 Artificial neural network 226 and computational time of
3.7 Comparisons of the test the deep learning models 234
accuracy and computational 4.6 Cross validation 236
time of the shallow-learning 5 Comparison of the performances
models 226 of the deep and shallow models 236
4 Deep learning models 228 6 Discussions 239
4.1 Variational autoencoder 7 Conclusions 240
assisted neural network 230 References 240

Nomenclature
ϕ SVR kernel function
θ coefficients
ANN artificial neural network
GAN generative adversarial network
kNNR k-nearest neighbor regressor

Machine Learning for Subsurface Characterization. https://doi.org/10.1016/B978-0-12-817736-5.00008-9
© 2020 Elsevier Inc. All rights reserved. 219