Oil and Gas

Participants will learn to:

1. Summarize the modes of machine learning.
2. Explain the key machine learning principles, such as dealing with inadequate data, model generalization, validating, and ensemble methods.
3. Process data through data cleaning, feature scaling, and transformation pipelines.
4. Apply supervised methods for classification and regression.
5. Develop a structured approach to unsupervised clustering and cluster evaluation.
6. Critically evaluate the various machine learning algorithms.

Modes of Machine Learning

• Supervised and unsupervised learning

Machine Learning Principles

• Dealing with poor or inadequate data
• Underfitting, overfitting and model generalization
• Testing and validating
• Ensemble methods

Data Processing

• Data cleaning, categorical variables
• Feature scaling and transformation pipelines

Supervised Machine Learning Models

• Linear, polynomial, logistic regression and regularized linear models
• Decision trees for classification and regression
• Ensemble learning, random forests, bagging and boosting for regression and classification
• Support vector machines for regression and classification
• Neural networks for regression and classification
• Introduction to deep neural networks and application to image segmentation
• Time-series analyses, ARMA, ARIMA and survival analyses

Unsupervised Machine Learning Models

• Dimensionality reduction
• Gaussian mixture models
• K-Means and hierarchical clustering
• DBSCAN

This course is aimed at subsurface data scientists, geoscientists, and engineers with some background in data-driven applications. Participants are expected to have experience of coding in Python and of machine learning or an introductory course on this topic.