3 Graphs in Machine Learning Application
This chapter covers:
- A detailed description of the role of graphs in the machine learning workflow
- A system for large scale graph processing
- How graphs are used to break down complex processing tasks
This chapter describes, more in detail, how graphs and machine learning can fit together helping to deliver better services to end users, data analysts or business people. In the previous 2 chapters, we introduced the general concepts about the machine learning project like:
- The different types of machine learning algorithms: supervised, unsupervised, semi-supervised
- The different phases that compose a generic machine learning projects: the 6 phases described in the CRISP-DM model such as business understanding, data understanding, data preparation, modelling, evaluation, deployment;
- The importance of the data quality and quantity to create a valuable and meaningful model that could provide accurate predictions;
- How to handle huge amount of data (big data) using a graph data model.
This chapter will walk you through how to combine the power of the graph model as a way for representing data that makes it easy to access and analyze as well as how to leverage the “intelligence” of the machine learning algorithms based on graph theory.
I would like to start this chapter with an image that represents the vision about the path of converting raw data in something that is more than “simple” knowledge or insight: it is wisdom.