nach oben

2023 | Buch

Kapitel lesen Erstes Kapitel lesen

Machine Learning for Earth Sciences

Using Python to Solve Geological Problems

verfasst von: Maurizio Petrelli

Verlag: Springer International Publishing

Buchreihe : Springer Textbooks in Earth Sciences, Geography and Environment

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

This textbook introduces the reader to Machine Learning (ML) applications in Earth Sciences. In detail, it starts by describing the basics of machine learning and its potentials in Earth Sciences to solve geological problems. It describes the main Python tools devoted to ML, the typival workflow of ML applications in Earth Sciences, and proceeds with reporting how ML algorithms work. The book provides many examples of ML application to Earth Sciences problems in many fields, such as the clustering and dimensionality reduction in petro-volcanological studies, the clustering of multi-spectral data, well-log data facies classification, and machine learning regression in petrology. Also, the book introduces the basics of parallel computing and how to scale ML models in the cloud. The book is devoted to Earth Scientists, at any level, from students to academics and professionals.

Inhaltsverzeichnis

Frontmatter

Basic Concepts of Machine Learning for Earth Scientists

Frontmatter

Chapter 1. Introduction to Machine Learning

Abstract

This chapter introduces the basics of machine learning to geologists. Toward this end, it first provides fundamental definitions and introduces common terminology. It then discusses the learning process and defines the different types of learning paradigms (i.e., supervised, unsupervised, and semisupervised).

Maurizio Petrelli

Chapter 2. Setting Up Your Python Environments for Machine Learning

Abstract

This chapter details how to prepare a Python environment to start working with Machine Learning in Earth Sciences. First, it shows how to set up a local Python environment, and then how to create a remote Linux instance. Finally, it explains how to start working with cloud-based machine learning environments.

Maurizio Petrelli

Chapter 3. Machine Learning Workflow

Abstract

This chapter describes machine learning workflows. It starts by introducing a typical five-step workflow made of (1) data acquisition, (2) pre-processing, (3) model training, (4) model validation, and (5) model deployment. Each step is described in detail and accompanied by Python examples.

Maurizio Petrelli

Unsupervised Learning

Frontmatter

Chapter 4. Unsupervised Machine Learning Methods

Abstract

This chapter introduces unsupervised machine learning methods. It starts by describing the algorithms for dimensionality reduction, which include principal component analysis and manifold learning. It then describes clustering methods, such as hierarchical clustering, DBSCAN, mean shift, K-means, spectral clustering, and Gaussian-mixture models.

Maurizio Petrelli

Chapter 5. Clustering and Dimensionality Reduction in Petrology

Abstract

Chapter 5 describes how to apply unsupervised machine learning methods in petrology. It focuses on analyzing the clinopyroxene erupted by Mt. Etna during the sequence of lava fountains that occurred between February and April of 2021. The application of clustering and dimensionality reduction techniques is described in detail.

Maurizio Petrelli

Chapter 6. Clustering of Multi-Spectral Data

Abstract

This chapter deals with the application of unsupervised machine learning methods to multi-spectral images deriving from Earth-observing satellite missions. It describes how to import, pre-process, describe, and analyze multi-spectral data that can be downloaded from access points such as USGS Earth Explorer, the Copernicus Open Access Hub, and Theia.

Maurizio Petrelli

Supervised Learning

Frontmatter

Chapter 7. Supervised Machine Learning Methods

Abstract

This chapter describes supervised learning methods for regression and classification tasks. They include Naive Bayes, Ridge Regression, Least Absolute Shrinkage and Selection Operator (LASSO), Elastic Net, Support Vector Machines, Supervised Nearest Neighbors, and Tree-Based Methods.

Maurizio Petrelli

Chapter 8. Classification of Well Log Data Facies by Machine Learning

Abstract

This chapter focuses on the classification by machine learning of facies in well-log data. It progressively develops a machine learning workflow that includes descriptive statistics, algorithm selection, model optimization, model training, and application to blind observations. Each step is discussed in detail.

Maurizio Petrelli

Chapter 9. Machine Learning Regression in Petrology

Abstract

This chapter applies machine-learning regression to petrology. It explains how to calibrate machine-learning thermo-barometers based on orthopyroxene crystals in equilibrium with the melt in a volcanic plumbing system. It also describes the calibration of a thermo-barometer based on orthopyroxenes crystals.

Maurizio Petrelli

Scaling Machine Learning Models

Frontmatter

Chapter 10. Parallel Computing and Scaling with Dask

Abstract

This chapter introduces basic concepts and definitions of parallel computing and model scaling. It starts by providing basic definitions and terminology and then introduces Daks, a Python library that provides object scalability to Python scientific libraries such as pandas, NumPy, and scikit-learn.

Maurizio Petrelli

Chapter 11. Scale Your Models in the Cloud

Abstract

This chapter shows how to scale machine-learning models in the cloud. In the context of cloud computing, the term “scaling” refers to the ability to quickly and efficiently change the capability of a computational resource to handle a model that no longer fits the available resources.

Maurizio Petrelli

Next Step: Deep Learning

Frontmatter

Chapter 12. Introduction to Deep Learning

Abstract

This chapter is about deep learning. It starts by introducing the basics of deep learning and then introduces PyTorch, a Python deep learning library. It also describes how to set up and train feedforward networks. Finally, it provides an example application dealing with deep learning potentials in the Earth Sciences.

Maurizio Petrelli

Titel: Machine Learning for Earth Sciences
verfasst von: Maurizio Petrelli
Verlag: Springer International Publishing
Electronic ISBN: 978-3-031-35114-3
Print ISBN: 978-3-031-35113-6
DOI: https://doi.org/10.1007/978-3-031-35114-3