Predictive Analytics using Networked Data in R

This course will show you how to use networked information to perform advanced predictive analysis. We will be discussing how to use the network's structure and information in a predictive way. We also introduce featurization which allows network features and other features to be added. This improves the performance of any analytic model. This course will show you how to use igraph to label and create a network of customers in a churn environment. The foundations of network learning will also be covered. Next, you will learn about heterophily and dyadicity and how they can help you gain important insights into your network. Next, you will use the igraph package functionality for computing network features. This includes both neighbor-based and node centric network features. The Google PageRank algorithm will be used to compute and validate network features. We will also demonstrate how to create a flat data set using the network, and then analyze it with logistic regression or random forest.