Call number:
9783030647773 (e-book)
In:
Water science and technology library, volume 99
Description / Table of Contents:
This book provides a step-by-step methodology and derivation of deep learning algorithms as Long Short-Term Memory (LSTM) and Convolution Neural Network (CNN), especially for estimating parameters, with back-propagation as well as examples with real datasets of hydrometeorology (e.g. streamflow and temperature) and environmental science (e.g. water quality).
Deep learning is known as part of machine learning methodology based on the artificial neural network. Increasing data availability and computing power enhance applications of deep learning to hydrometeorological and environmental fields. However, books that specifically focus on applications to these fields are limited.
Most of deep learning books demonstrate theoretical backgrounds and mathematics. However, examples with real data and step-by-step explanations to understand the algorithms in hydrometeorology and environmental science are very rare.
This book focuses on the explanation of deep learning techniques and their applications to hydrometeorological and environmental studies with real hydrological and environmental data. This book covers the major deep learning algorithms as Long Short-Term Memory (LSTM) and Convolution Neural Network (CNN) as well as the conventional artificial neural network model.
Type of Medium:
12
Pages:
1 Online-Ressource (xiv, 204 Seiten)
,
Illustrationen, Diagramme, Karten
ISBN:
9783030647773
,
978-3-030-64777-3
ISSN:
0921-092X
,
1872-4663
Series Statement:
Water science and technology library volume 99
URL:
Ebook (access only within the AWI network)
DOI:
10.1007/978-3-030-64777-3
Language:
English
Note:
Contents
1 Introduction
1.1 What is Deep Learning?
1.2 Pros and Cons of Deep Learning
1.3 Recent Applications of Deep Learning in Hydrometeorological and Environmental Studies
1.4 Organization of Chapters
1.5 Summary and Conclusion
References
2 Mathematical Background
2.1 Linear Regression Model
2.1.1 Simple Linear Regression
2.1.2 Multiple Linear Regression
2.2 Time Series Model
2.2.1 Autoregressive Model (AR)
2.3 Probability Distributions
2.3.1 Normal Distributions
2.3.2 Gamma Distribution
2.4 Exercises
References
3 Data Preprocessing
3.1 Normalization
3.2 Data Splitting for Training and Testing
3.3 Exercises
4 Neural Network
4.1 Terminology in Neural Network
4.1.1 Components of Neural Network
4.1.2 Activation Functions
4.1.3 Error and Loss Function
4.1.4 Softmax and One-Hot Encoding
4.2 Artificial Neural Network
4.2.1 Simplest Network
4.2.2 Feedforward and Backward Propagation
4.2.3 Network with Multiple Input and Output Variables
4.2.4 Python Coding of the Simple Network
4.3 Exercises
5 Training a Neural Network
5.1 Initialization
5.2 Gradient Descent
5.3 Backpropagation
5.3.1 Simple Network
5.3.2 Full Neural Network
5.3.3 Python Coding of Network
5.4 Exercises
Reference
6 Updating Weights
6.1 Momentum
6.2 Adagrad
6.3 RMSprop
6.4 Adam
6.5 Nadam
6.6 Python Coding of Updating Weights
6.7 Exercises
References
7 Improving Model Performance
7.1 Batching and Minibatch
7.2 Validation
7.2.1 Python Coding of K-Fold Cross-Validation
7.3 Regularization
7.3.1 L-Norm Regularization
7.3.2 Dropout
7.3.3 Python Coding of Regularization
7.4 Exercises
Reference
8 Advanced Neural Network Algorithms
8.1 Extreme Learning Machine (ELM)
8.1.1 Basic ELM
8.1.2 Generalized ELM
8.1.3 Python Coding
8.2 Autoencoder
8.2.1 Vanilla Autoencoder
8.2.2 Regularized Autoencoder
8.2.3 Python Coding of Regularized AE
8.3 Exercises
Reference
9 Deep Learning for Time Series
9.1 Recurrent Neural Network
9.1.1 Backpropagation
9.1.2 Backpropagation Through Time (BPTT)
9.2 Long Short-Term Memory (LSTM)
9.2.1 Basics of LSTM
9.2.2 Example of LSTM
9.2.3 Backpropagation of a Simple LSTM
9.2.4 Backpropagation Through Time (BPTT)
9.3 Gated Recurrent Unit (GRU)
9.3.1 Basics of GRU
9.3.2 Example of GRU
9.3.3 Backpropagation of a Simple GRU Model
9.4 Exercises
References
10 Deep Learning for Spatial Datasets
10.1 Convolutional Neural Network (CNN)
10.1.1 Definition of Convolution
10.1.2 Elements of CNN
10.2 Backpropagation of CNN
10.3 Exercises
11 Tensorflow and Keras Programming for Deep Learning
11.1 Basic Keras Modeling
11.2 Temporal Deep Learning (LSTM and GRU)
11.3 Spatial Deep Learning (CNN)
11.4 Exercises
References
12 Hydrometeorological Applications of Deep Learning
12.1 Stochastic Simulation with LSTM
12.1.1 Mathematical Description for Stochastic Simulation with LSTM
12.1.2 Colorado Monthly Streamflow
12.1.3 Results of Colorado River
12.1.4 Python Coding
12.1.5 Matlab Coding
12.2 Forecasting Daily Temperature with LSTM
12.2.1 Preparing the Data
12.2.2 Methodology
12.2.3 Results
12.2.4 Python Coding
12.3 Exercises
References
13 Environmental Applications of Deep Learning
13.1 Remote Sensing of Water Quality Using CNN
13.1.1 Introduction
13.1.2 Study Area and Monitoring
13.1.3 Field Data Collection
13.1.4 Point-Centered Regression CNN (PRCNN)
13.1.5 Results and Discussion
13.1.6 Conclusion
13.1.7 Python Coding
References
Permalink
