In order to create the ensemble, we will utilize the openensembles library that we presented in Chapter 8, Clustering. As our dataset does not contain labels, we cannot use the homogeneity score in order to evaluate our clustering models. Instead, we will use the silhouette score, which evaluates how cohesive each cluster is and how separate different clusters are. First, we must load our dataset, which is provided in the WHR.csv file. The second file that we load, Regions.csv, contains the region that each country belongs to. We will utilize the data from 2017, as 2018 has a lot of missing data (for example, Delivery quality and Democratic quality are completely absent). We will fill any missing data using the median ...
Creating the ensemble
Get Hands-On Ensemble Learning with Python now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.