Book Image

Network Science with Python

By : David Knickerbocker
Book Image

Network Science with Python

By: David Knickerbocker

Overview of this book

Network analysis is often taught with tiny or toy data sets, leaving you with a limited scope of learning and practical usage. Network Science with Python helps you extract relevant data, draw conclusions and build networks using industry-standard – practical data sets. You’ll begin by learning the basics of natural language processing, network science, and social network analysis, then move on to programmatically building and analyzing networks. You’ll get a hands-on understanding of the data source, data extraction, interaction with it, and drawing insights from it. This is a hands-on book with theory grounding, specific technical, and mathematical details for future reference. As you progress, you’ll learn to construct and clean networks, conduct network analysis, egocentric network analysis, community detection, and use network data with machine learning. You’ll also explore network analysis concepts, from basics to an advanced level. By the end of the book, you’ll be able to identify network data and use it to extract unconventional insights to comprehend the complex world around you.
Table of Contents (17 chapters)
1
Part 1: Getting Started with Natural Language Processing and Networks
5
Part 2: Graph Construction and Cleanup
9
Part 3: Network Science and Social Network Analysis

Introducing ML

ML is a set of techniques that enable computers to learn from patterns and behavior in data. It is often said that there are three different kinds of ML: Supervised, Unsupervised, and Reinforcement learning.

In supervised ML, an answer – called a label – is provided with the data to allow for an ML model to learn the patterns that will allow it to predict the correct answer. To put it simply, you give the model data and an answer, and it figures out how to predict correctly.

In unsupervised ML, no answer is provided to the model. The goal is usually to find clusters of similar pieces of data. For instance, you could use clustering to identify the different types of news articles present in a dataset of news articles, or to find topics that exist in a corpus of text. This is similar to what we have done with community detection.

In reinforcement learning, a model is given a goal and it gradually learns how to get to this goal. In many reinforcement...