Page 156 - Intelligent Digital Oil And Gas Fields

Components of Artificial Intelligence and Data Analytics 119

Table 4.4 Summary of the Most Popular Methods and Classifiers in ML, as
Representative of the Three Main Paradigms Described Previously, With Some
Suggested References for Further Reading
ML Family
Type of Suggested
ML Paradigm Problem Technique Reading

Supervised Regression Linear regression (LR) Hastie et al.
learning • Ordinary least squares (2011)
• Stepwise and moving LR James et al.
Penalized LR (2014)
• Ridge LR Leskovec
• Elastic nets et al. (2014)
Nonlinear regression Brownlee
• Multivariate adaptive regres- (2014)
sion splines (MARS)
• Support vector machine
(SVM)
• K-nearest neighbor
• Neural network (NN)
Decision trees for regression
• Classification and regression
trees (CART)
• Conditional decision trees
• Bagging CART
• Random forest (RF)
• Gradient boosted machine
(GBM)
Classification Linear classification
• Logistic regression
• Discriminant analysis
Nonlinear classification
• Mixture, regularized, qua-
dratic and flexible discrimi-
nant analysis
• Support vector machine
(SVM)
• K-nearest neighbor
• Naive Bayes
Nonlinear classification with
decision trees
• Classification and regression
trees (CART)
• Bootstrapped aggregation
(Bagging) CART
• Random forest (RF)
• Gradient boosted machine
(GBM)
Continued