OPTICS Ordering points to identify the clustering structure

No account yet ?

Point Order to Identify Clustering Structure is an algorithm for finding density-based clusters in spatial data.

OPTICS is similar to DBSCAN, but it addresses one of the main weaknesses of DBSCAN, namely the problem of detecting significant clusters in data of varying density.

To do this, the points in the database are ordered such that the spatially closest points become neighbors in order.

A special distance is stored for each point that represents the density that must be accepted for a cluster in order for both points to belong to it.

This is represented by a dendrogram.

Anomaly detection with OPTICS

Tested in Anaconda and Python 3.7

from sklearn.cluster import OPTICS
from sklearn.datasets import make_blobs
from numpy import quantile, where, random
import matplotlib.pyplot as plt
 
random.seed(123)
x, _ = make_blobs(n_samples=350, centers=1, cluster_std=.4, center_box=(20, 5))
 
plt.scatter(x[:,0], x[:,1])
plt.grid(True)
plt.show() 
 
model = OPTICS().fit(x)
print(model)
 
scores = model.core_distances_
 
thresh = quantile(scores, .98)
print(thresh) 
 
index = where(scores >= thresh)
values = x[index]
print(values)
 
plt.scatter(x[:,0], x[:,1])
plt.scatter(values[:,0],values[:,1], color='r')
plt.legend(("normal", "anomal"), loc="best", fancybox=True, shadow=True)
plt.grid(True)
plt.show()
 

Source : https://www.datatechnotes.com/2020/12/anomaly-detection-with-optics-in-python.html