packtpublishing / machine-learning-for-time-series-with-python

Machine Learning for Time-Series with Python.Published by Packt

Machine-Learning-for-Time-Series-with-Python

Become proficient in deriving insights from time-series data and analyzing a model’s performance

Links

• Amazon
• Packt Publishing

Key Features

• Explore popular and modern machine learning methods including the latest online and deep learning algorithms
• Learn to increase the accuracy of your predictions by matching the right model with the right problem
• Master time-series via real-world case studies on operations management, digital marketing, finance, and healthcare

What you will learn

• Understand the main classes of time-series and learn how to detect outliers and patterns
• Choose the right method to solve time-series problems
• Characterize seasonal and correlation patterns through autocorrelation and statistical techniques
• Get to grips with time-series data visualization
• Understand classical time-series models like ARMA and ARIMA
• Implement deep learning models, like Gaussian processes, transformers, and state-of-the-art machine learning models
• Become familiar with many libraries like Prophet, XGboost, and TensorFlow

Who This Book Is For

This book is ideal for data analysts, data scientists, and Python developers who are looking to perform time-series analysis to effectively predict outcomes. Basic knowledge of the Python language is essential. Familiarity with statistics is desirable.

Table of Contents

1. Introduction to Time-Series with Python
2. Time-Series Analysis with Python
3. Preprocessing Time-Series
4. Introduction to Machine Learning for Time-Series
5. Forecasting with Moving Averages and Autoregressive Models
6. Unsupervised Methods for Time-Series
7. Machine Learning Models for Time-Series
8. Online Learning for Time-Series
9. Probabilistic Models for Time-Series
10. Deep Learning for Time-Series
11. Reinforcement Learning for Time-Series
12. Multivariate Forecasting

Author Notes

I've heard from a few people struggling with tsfresh and featuretools for chapter 3.

My PR for tsfresh was merged mid-December fixing a version incompatibility - featuretools went through many breaking changes with the release of version 1.0.0 (congratulations to the team!). Please see how to fix any problems in the discussion here.

Download a free PDF

If you have already purchased a print or Kindle version of this book, you can get a DRM-free PDF version at no cost.
Simply click on the link to claim your free PDF.

https://packt.link/free-ebook/9781801819626

machine-learning-for-time-series-with-python's Issues

Chapter 3 Preprocessing notebook

Hello,
In the Chapter 3 Preprocessing notebook there seem to be several errors in the code.

When installing 'featuretools[tsfresh]' in colab the following error occurs:
WARNING: featuretools 0.27.1 does not provide the extra 'tsfresh'

When running the first code bracket using feature tools in colab the following error occurs:
TypeError: import_optional_dependency() got an unexpected keyword argument 'errors'

It appears colab installs featuretools version 0.27.1
If I install the most recent version of featuretoolls (i.e., 1.3.0) and run the code bracket below Automated Feature Extraction I get the following error:
AttributeError: 'EntitySet' object has no attribute 'entity_from_dataframe'

If it is versioning that is an issue, please let me know what version of featuretools the code in the book was based on. If it is not versioning, would you be able to advise what the issue is?

Let me know if you have any questions. Thanks!

Best,
Nils

this repo's minor readme issue

Hi Ben!

Not a book issue, but one with the readme associated with this repo. The Key Features section starts fine but in the middle it accidentally switches into a description of another Pack book on Power BI. Looking forward to reading your book!

Alex

Time Series Clustering

The coverage of time series clustering is poor. Actually does not exist beyond referencing a few libraries which people can find themselves. Any follow up with examples here would help I think.

the code is not complete

Hi Ben,

I have high expectation on your book and am very excited to learn. However, the codes published on Github is incomplete. For example, in chapter 2, where are the codes from page 52 to 53 especially page 53? I am very interested to know how you did the code in page 53. Also some codes in jupyter notebook are not explained either in notebook and book. I am kind of disappointed. Could you please republish your complete jupyter notebook? Thanks

Hard to believe these are typos

This is from the documentation of ARCH package
"The null hypothesis of the Augmented Dickey-Fuller is that there is a unit root, with the alternative that there is no unit root. If the p-value is above a critical size, then the null cannot be rejected that there and the series appears to be a unit root."

on p. 152 a p-value of 0.997 is reported (in a figure which references KPSS test while the table shown clearly says that it is an ADF test) and the following conclusion is made "Given the p-value of 0.997, we can reject our null hypothesis of the unit root, and we conclude that our process is weakly stationary."

In a couple of pages "Please note that we need to set trend="t" here so that the model includes a constant.
If not, we would get a spurious regression." In the statsmodels' ARIMA class documentation, it is stated clearly that trend is "Parameter controlling the deterministic trend. Can be specified as a string where ‘c’ indicates a constant term, ‘t’ indicates a linear trend in time, and ‘ct’ includes both."

It's quite doubtful that the author of this book understands what he is writing about. But at least he could have checked the documentation of the packages used in it.

How to use ta for Trading_with_DQN?

Any example of using ta for features in Trading_with_DQN notebook?