Implementing DBSCAN clustering

DBSCAN is a very flexible approach to clustering. We just need to specify a value for ɛ, also referred to as eps. As we have discussed, the ɛ value determines the size of the ɛ-neighborhood around an instance. The minimum samples hyperparameter indicates how many instances around an instance are needed for it to be considered a core instance.

Note

We use DBSCAN to cluster the same income gap data that we worked with in the previous section.

Let’s build a DBSCAN clustering model:

1. We start by loading familiar libraries, plus the DBSCAN module:

   import pandas as pd
   from sklearn.preprocessing import MinMaxScaler
   from sklearn.pipeline import make_pipeline
   from sklearn.cluster import DBSCAN
   from sklearn.impute import KNNImputer
   from sklearn.metrics import silhouette_score
   import matplotlib.pyplot as plt
   import os
   import sys
   sys.path.append(os.getcwd() + "/helperfunctions")

2. We import the code to load and...