Graph-Powered Machine Learning
Publisher Finelybook 出版社：Manning Publications; 1st edition (29 Nov. 2021)
pages 页数：496 pages
Thanks for purchasing the MEAP of Graph Powered Machine Learning. Graph-based machine learning is becoming a very important trend in Artificial Intelligence, transcending a lot of other techniques. Google, Facebook, and E-bay – to cite some of them – have multiple projects involving graphs, and more specifically graph models and graph algorithms, as empowering mechanism behind the most advanced services they are providing to their end users.
Graph-Powered Machine Learning is a practical guide to effectively using graphs in machine learning applications, driving you in all the stages necessary for building complete solutions where graphs play a key role. It focuses on methods, algorithms, and design patterns related to graphs. Based on my personal experience on building complex machine learning applications, this book suggests many recipes in which graphs are the main ingredient to prepare a tasty product to your customers. Across the lifecycle of a machine learning project such approaches can be useful under several aspects:managing data sources more efficiently, implement better algorithms, storing the prediction model so that they can be accessed faster, and visualizing the results in a more effective way for further analysis.
The book is divided into three parts. The first part is introductory to the topic. The three chapters introduce main graph and machine learning concepts from the basics. Furthermore, the role of graphs in Big Data Platforms and Machine learning is highlighted and presented using multiple scenarios.
The second part is the core of the book. Several techniques are described, from data source modeling, to algorithm design both leveraging graphs as underlying technology. A lot of optimization approaches, best practices, and common pitfalls are detailed to help data scientists or data engineers define the infrastructure and choose the right approaches since the beginning of their projects.
The last part presents three different applications. Here concrete end-to-end projects are discussed and for each of them the architecture, design best practice, and common pitfalls will be illustrated.
I hope that what you’ll get access to will be of interest for your current machine learning project and for your learning path as a data scientist, a data engineer or a data architect, and that it will occupy an important place on your digital or, even better, physical bookshelf.
Please be sure to stop by the liveBook Discussion Forum with any feedback you have. With your help, I’m sure the final book will be great!
1 Machine Learning and Graph:An introduction
2 Graph Data Engineering
3 Graphs in Machine Learning Application
4 Content-Based Recommendation
5 Collaborative Filtering
6 Session-Based Recommendation
7 Context-Aware and Hybrid Recommendation
8 Fighting fraud:Introduction and basic tools
9 Fighting fraud:Proximity-based algorithms
10 Fighting fraud:Social network analysis
11 Taming text with graphs
12 Knowledge graphs