Clustering Analyzer

What it does

The Clustering Analyzer skill enables AI agents to identify patterns and groupings within data clusters using various algorithms. It supports K-Means, DBSCAN, and Hierarchical clustering techniques, providing visualization tools like plots and dendrograms for analysis. The skill also offers evaluation metrics such as silhouette scores and cluster statistics to assess the quality of the resulting clusters.

When to use it

• To segment customers based on purchasing behavior from a CSV file.
• To identify distinct groups within a dataset of product reviews.
• To explore potential groupings in sensor data for anomaly detection.
• To determine the optimal number of clusters using the elbow method with K-Means.

Key capabilities

• K-Means Clustering: Partitioning data into k clusters, utilizing the elbow method to find an appropriate 'k'.
• DBSCAN Clustering: Density-based clustering suitable for identifying clusters with arbitrary shapes.
• Hierarchical Clustering: Creates a hierarchy of clusters represented by dendrograms.
• Evaluation Metrics: Calculates silhouette scores and cluster statistics to evaluate cluster quality.
• Visualization: Generates 2D/3D plots, dendrograms, and elbow curves for data exploration.
• Data Export: Exports labeled data and cluster summaries to CSV files.

Example prompts

• "Analyze the 'customers.csv' file using K-Means clustering with 3 clusters."
• "Perform DBSCAN clustering on the 'data.csv' file with eps=0.5 and min_samples=5."
• "Find the optimal number of clusters for the 'data.csv' file using the elbow method with K-Means."

Tips & gotchas

• The skill requires a CSV file as input data; specify columns if needed during loading.
• DBSCAN parameters like eps and min_samples significantly impact results – experiment to find optimal values.
• Familiarity with clustering algorithms (K-Means, DBSCAN, Hierarchical) is helpful for interpreting the results.