基于Python的DBSCAN聚类：为鸢尾花数据集添加DBSCAN聚类列并导出

Got it, let's walk through exactly how to add that DBSCAN column to your dataset and export the updated file. Here's a step-by-step breakdown with code tailored to your specific needs:

Step 1: Import Required Libraries #

We’ll need pandas for handling CSV data and DBSCAN from scikit-learn for clustering:

import pandas as pd
from sklearn.cluster import DBSCAN

Step 2: Load Your Dataset #

Use a raw string for your Windows file path (the r prefix) to avoid backslash-related errors:

# Load the existing CSV file
file_path = r"C:\Users\05807\Desktop\New folder (5)\FinalX.csv"
df = pd.read_csv(file_path)

Step 3: Prepare Numerical Features for DBSCAN #

DBSCAN only works with numerical data, so we’ll select the four iris measurement columns as our clustering inputs:

# Pick numerical columns for clustering (exclude text/ID/existing cluster columns)
numerical_features = df[["Sepal_length", "Sepal_width", "Petal_length", "Petal_width"]]

Step 4: Run DBSCAN and Format Cluster Labels #

By default, DBSCAN returns integer labels (0, 1, 2, etc., plus -1 for noise points). We’ll convert these to the "Cluster X" format you want, and clearly mark noise:

# Initialize and fit DBSCAN (tweak eps/min_samples to get your desired 3 clusters)
dbscan = DBSCAN(eps=0.5, min_samples=5)
cluster_labels = dbscan.fit_predict(numerical_features)

# Convert labels to human-readable format
formatted_labels = []
for label in cluster_labels:
    if label == -1:
        formatted_labels.append("Noise")
    else:
        formatted_labels.append(f"Cluster {label + 1}")  # Start counting from 1 instead of 0

Note: Adjust eps (maximum distance between points in a cluster) and min_samples (minimum points required to form a cluster) if you don’t get 3 distinct clusters initially—test small changes to find the best fit for your data.

Step 5: Insert the DBSCAN Column After "Hierarchical" #

We’ll place the new column exactly where you want it:

# Find the position of the "Hierarchical" column
hierarchical_col_index = df.columns.get_loc("Hierarchical")
# Insert the DBSCAN column right after it
df.insert(hierarchical_col_index + 1, "DBSCAN", formatted_labels)

Step 6: Export the Updated Dataset #

Save the modified data to a new CSV file:

# Define output path (adjust this to your preferred location/filename)
output_path = r"C:\Users\05807\Desktop\New folder (5)\FinalX_With_DBSCAN.csv"
df.to_csv(output_path, index=False)

print(f"Updated dataset saved successfully to: {output_path}")

Full Working Code #

Putting it all together, here’s the complete script you can run:

import pandas as pd
from sklearn.cluster import DBSCAN

# Load original dataset
file_path = r"C:\Users\05807\Desktop\New folder (5)\FinalX.csv"
df = pd.read_csv(file_path)

# Select numerical features for clustering
numerical_features = df[["Sepal_length", "Sepal_width", "Petal_length", "Petal_width"]]

# Run DBSCAN clustering
dbscan = DBSCAN(eps=0.5, min_samples=5)
cluster_labels = dbscan.fit_predict(numerical_features)

# Format labels to "Cluster X" style
formatted_labels = []
for label in cluster_labels:
    if label == -1:
        formatted_labels.append("Noise")
    else:
        formatted_labels.append(f"Cluster {label + 1}")

# Insert DBSCAN column after Hierarchical
hierarchical_col_index = df.columns.get_loc("Hierarchical")
df.insert(hierarchical_col_index + 1, "DBSCAN", formatted_labels)

# Export updated data to new CSV
output_path = r"C:\Users\05807\Desktop\New folder (5)\FinalX_With_DBSCAN.csv"
df.to_csv(output_path, index=False)

print("Done! Your dataset now includes the DBSCAN cluster column.")

Just execute this script, and you’ll have your new CSV with the DBSCAN column positioned exactly after "Hierarchical".

内容的提问来源于stack exchange，提问作者Snyder Fox