Challenge: Using DBSCAN Clustering to Detect Outliers

Завдання

Swipe to start coding

Now, you will apply the DBSCAN clustering algorithm to detect outliers on a simple Iris dataset.
You have to:

Specify the parameters of the DBScan algorithm: set eps equal to 0.35 and min_samples equal to 6.
Fit the algorithm and provide clustering.
Get outlier indexes and indexes of normal data. Pay attention that outliers detected by the algorithm have a -1 label.

Рішення

import pandas as pd

import matplotlib.pyplot as plt

from sklearn.cluster import DBSCAN

from sklearn import datasets

import numpy as np

# Load the Iris dataset

iris = datasets.load_iris()

# Create a DataFrame directly from the data and target arrays

data = pd.DataFrame(data=np.c_[iris.data, iris.target], columns=iris.feature_names + ['target'])

# Extract the features (Sepal Length, Sepal Width, Petal Length, Petal Width)

features = data.iloc[:, :2] # We only consider the first two features for visualization

# Apply DBSCAN clustering for outlier detection

dbscan = DBSCAN(eps=0.35, min_samples=6)

dbscan_labels = dbscan.fit_predict(features)

# Visualize the results (2D scatter plot)

outlier_indices = np.where(dbscan_labels == -1)

normal_indices = np.where(dbscan_labels != -1)

plt.scatter(features.loc[normal_indices, 'sepal length (cm)'], features.loc[normal_indices, 'sepal width (cm)'],

c='blue', label='Normal Data', alpha=0.6)

plt.scatter(features.loc[outlier_indices, 'sepal length (cm)'], features.loc[outlier_indices, 'sepal width (cm)'],

c='red', marker='x', label='Outliers', s=100)

plt.title('DBSCAN Outlier Detection on Iris Dataset')

plt.xlabel('Sepal Length (cm)')

plt.ylabel('Sepal Width (cm)')

plt.legend()

plt.show()

# Identify and print the number of outliers (anomalies) detected

outliers = data.iloc[outlier_indices]

print(f"Number of outliers detected: {len(outliers)}")

Все було зрозуміло?

Дякуємо за ваш відгук!

Секція 3. Розділ 2