![](https://csdnimg.cn/release/download_crawler_static/87805726/bg8.jpg)
presented in great details, and multiple graph-based algorithms are described. The most
complex concepts are illustrated with concrete scenarios and concrete applications are
designed.
This writing aims to be a concrete guide that would help you to successfully install a
working application in the production. Hence, optimization techniques and heuristics are also
described to help you to deal with real data, real problem, real users. Not just toy examples
are discussed, but also end-to-end applications from real world use cases are depicted and
illustrated with some suggestions to deal with concrete problem.
If these arguments solicited your interest this is definitively the right book for you.
WHO SHOULDN’T READ THIS BOOK
If you are looking for a book with a list of basic machine learning methods or if you are looking
for a high/business level introduction to machine learning techniques may be this is not the
right book for you. If you are not interested in graphs, definitely leave it on the shelf!
1.1 Introduction to Machine Learning
1.1.1 Machine Learning Project Lifecycle
A machine learning project is a human process as well as a software project. It involves a
large quantity of people, a lot of communication, a lot of work and a mixed of skills. It requires
a well-defined approach to be effective. We start our long journey, defining a workflow with
clear steps/components that is used along all the book. The mental schema proposed, that is
just one of the possible schemas, helps also to better understand the role of the graphs in the
development and deployment of a successful machine learning project.
Delivering machine learning solutions is a very complex process and requires more than
just selecting the right algorithm(s). It involves multiple tools, a lot of data and different
people. Such projects involve numerous tasks related to:
• selecting the data sources;
• gathering;
• understanding;
• cleaning;
• processing;
• evaluating the results;
• deploying.
After deployment, it is necessary to monitor the application and tune it.
One of the most commonly used processes for data mining projects is CRoss Industry
Standard Process for Data Mining [With and Hipp (2000)]. Although the CRISP-DM model
was designed for Data Mining it can also be applied to generic machine learning project. Key
features of the CRISP-DM that make it attractive as part of the base workflow model are:
• it is not-proprietary;
• it is application, industry, and tool neutral;
• it explicitly views the data analytics process from both application-focused and a
©Manning Publications Co. We welcome reader comments about anything in the manuscript - other than typos and
other simple mistakes. These will be cleaned up during production of the book by copyeditors and proofreaders.
https://forums.manning.com/forums/graph-powered-machine-learning