Book Image

Building Machine Learning Systems with Python

Book Image

Building Machine Learning Systems with Python

Overview of this book

Machine learning, the field of building systems that learn from data, is exploding on the Web and elsewhere. Python is a wonderful language in which to develop machine learning applications. As a dynamic language, it allows for fast exploration and experimentation and an increasing number of machine learning libraries are developed for Python.Building Machine Learning system with Python shows you exactly how to find patterns through raw data. The book starts by brushing up on your Python ML knowledge and introducing libraries, and then moves on to more serious projects on datasets, Modelling, Recommendations, improving recommendations through examples and sailing through sound and image processing in detail. Using open-source tools and libraries, readers will learn how to apply methods to text, images, and sounds. You will also learn how to evaluate, compare, and choose machine learning techniques. Written for Python programmers, Building Machine Learning Systems with Python teaches you how to use open-source libraries to solve real problems with machine learning. The book is based on real-world examples that the user can build on. Readers will learn how to write programs that classify the quality of StackOverflow answers or whether a music file is Jazz or Metal. They will learn regression, which is demonstrated on how to recommend movies to users. Advanced topics such as topic modeling (finding a text's most important topics), basket analysis, and cloud computing are covered as well as many other interesting aspects.Building Machine Learning Systems with Python will give you the tools and understanding required to build your own systems, which are tailored to solve your problems.
Table of Contents (20 chapters)
Building Machine Learning Systems with Python
Credits
About the Authors
About the Reviewers
www.PacktPub.com
Preface
Index

Penalized regression


The important variations of OLS regression fall under the theme of penalized regression. In ordinary regression, the returned fit is the best fit on the training data, which can lead to overfitting. Penalizing means that we add a penalty for overconfidence in the parameter values.

Tip

Penalized regression is about tradeoffs

Penalized regression is another example of the bias-variance tradeoff. When using a penalty, we get a worse fit in the training data as we are adding bias. On the other hand, we reduce the variance and tend to avoid overfitting. Therefore, the overall result might be generalized in a better way.

L1 and L2 penalties

There are two types of penalties that are typically used for regression: L1 and L2 penalties. The L1 penalty means that we penalize the regression by the sum of the absolute values of the coefficients, and the L2 penalty penalizes by the sum of squares.

Let us now explore these ideas formally. The OLS optimization is given as follows:

In the...