Time series forecasting using deep learning : combining pytorch, RNN, TCN, and deep neural network models to provide production-ready prediction solutions / Ivan Gridin.

Format:

Book

Author/Creator:

Gridin, Ivan, author.

Language:

English

Subjects (All):

Deep learning (Machine learning).

Neural networks (Computer science).

Python (Computer program language).

Time-series analysis--Data processing.

Time-series analysis.

Physical Description:

1 online resource (314 pages) : illustrations

Edition:

First edition.

Place of Publication:

India : BPB Publications, [2022]

Summary:

This book is written for engineers, data scientists, and stock traders who want to build time series forecasting programs using deep learning. Possessing some familiarity of Python is sufficient, while a basic understanding of machine learning is desirable but not needed. -- Edited summary from book.

Contents:

Intro

Cover Page

Title Page

Copyright Page

About the Author

About the Reviewer

Acknowledgement

Preface

Errata

Table of Contents

1. Time Series Problems and Challenges

Structure

Objectives

Introduction to time series analysis and time series forecasting

Time series analysis

Time series forecasting

Time series characteristics

Random walk

Import part

Random walk generation

Trend

Result

Seasonality

Stationarity

Time series common problems

Forecasting

Modelling

Anomaly detection

Classical approaches

Autoregressive model (AR)

Autoregressive integrated moving average model

Seasonal autoregressive integrated moving average

Holt Winter's exponential smoothing

Classical approaches: Pros and cons

Promise of Deep Learning

Python for time series analysis

Pandas

Numpy

Matplotlib

Statmodels

Scikit-learn

PyTorch

Conclusion

Points to remember

Multiple choice questions

Answers

Key terms

2. Deep Learning with PyTorch

Setting up PyTorch

PyTorch as derivative calculator

Function creation

Computing function value

Create computational graph

PyTorch basics

Tensors

Tensor creation

Random tensor

Reproducibility

Common tensor types

Tensor methods and attributes

Math functions

Deep Learning layers

Linear layer

Convolution

Kernel

Weight

Padding

Stride

Pooling

Dropout

Activations

ReLU

Sigmoid

Tanh

Neural network architecture

Improving neural network performance.

Do not put two same layers in a row

Prefer ReLU activation at first

Start from fully connected network

More layers are better than more neurons

Use dropout

Put Deep Learning blocks in the beginning

Training

Loss functions

Absolute loss

Mean squared error

Smooth L1 loss

Optimizers

Adagrad

Adadelta

Adam

Stochastic Gradient Descent (SGD)

Time series forecasting example

Train, validation and test datasets

3. Time Series as Deep Learning Problem

Problem statement

Regression versus classification

Time series regression problems

Time series classification problems

Univariate versus multivariate

Univariate input - univariate output

Multivariate input - univariate output

Multivariate input - multivariate output

Many-to-many

Many-to-one

Single-step versus multi-step

Single-step

Multi-step

Single multi-step model

Multiple single-step model

Recurrent single-step model

Datasets

Feature engineering

Time series pre-processing and post-processing

Normalization

Trend removal

Differencing

Sliding window

Effectiveness and loss function

Static versus dynamic

Architecture design

Training, validating and testing

Alternative model

Model optimization

Summary

Example: UK minimal temperature prediction problem

Dataset

Architecture

Testing

Making script reproducible

Number of features

Preparing datasets

Initializing models

Loss function and optimization algorithm

Training process

Evaluation on test set

Getting results

Points to remember.

Multiple choice questions

4. Recurrent Neural Networks

Recurrent neural network

Making this script reproducible

Parameters

Preparing datasets for training

Initializing the model

Evaluation

Performance on test dataset

Training progress

Gated recurrent unit

Long short-term memory

5. Advanced Forecasting Models

Encoder-decoder model

Encoder-decoder training

Recursive

Teacher forcing

Mixed teacher forcing

Implementing the encoder-decoder model

Encoder layer

Decoder layer

Encoder-decoder model class

Model evaluation

Example

Global parameters

Generating datasets

Initializing Encoder-decoder model

Prediction

Visualizing results

Temporal convolutional network

Casual convolution

Dilation

Temporal convolutional network design

Implementing the temporal convolutional network

Crop layer

Temporal casual layer

Implementing temporal convolutional network

TCN prediction model

Generating time series

Preprocessing

Initializing the model.

Defining optimizer and loss function

Performance on the test dataset

Answer

6. PyTorch Model Tuning with Neural Network Intelligence

Objective

Neural Network Intelligence framework

Hyper-parameter tuning

Search space

Trial

Tuner

Hyper-parameter tuning in action

NNI Quick Start

Defining search space

Search configuration

NNI API

NNI search space

NNI Trial Integration

Time series model hyper-parameter tuning example

Deep Learning model trial

Dataset, optimizer, and model initialization

NNI search

Maximum number of trials

Neural Architecture Search

Hybrid models

Implementing hybrid model

Casual convolution layer

Hybrid model

Optional casual convolution layer

Obligatory RNN layer

Optional fully connected layer

Hybrid model trial

Hybrid model search space

Hybrid model architecture search

7. Applying Deep Learning to Real-world Forecasting Problems

Rain prediction

Model preparation function

Model hyper-parameters

Locations and features to train on

Sliding window dataset

Train-validation split

Converting all datasets to tensors:

Optimizer

Loss function

Loading the trained model

Making the predictions

Alternative predictions

Computing scores

Printing the results.

COVID-19 confirmed cases forecast

Preparing sliding window datasets

Creating train/validation datasets

Converting datasets to tensors

Training and getting the results

Creating the input

Making the prediction

Plotting the prediction

Algorithmic trading

Preparing sliding window dataset

Creating train and validation datasets

Preparing tensors

Model initializing

Best hyper-parameters

Creating tensors

Initializing and loading the model

Evaluating

8. PyTorch Forecasting Package

Introduction to PyTorch Forecasting package

Working with TimeSeriesDataset

Creating TimeSeriesDataSet

Working with TimeSeriesDataSet object

Initializing built-in PyTorch Forecasting model

Initializing Deep Autoregressive model

Creating custom PyTorch Forecasting model

Defining PyTorch Forecasting model

Implementing the forward method

Initializing the custom model

A complete example

9. What is Next?

Classical time series analysis

Deep learning

Studying the best solutions

Do not be afraid of science

Expanding your toolbox

Index.

Notes:

Description based on online resource; title from PDF title page (BPB Publications, viewed February 13, 2023).

Description based on print version record.

ISBN:

9789391392659

9391392652

OCLC:

1380465188