Non-Hierarchical Clustering

When data must be segmented or grouped, K-Means clustering provides a straightforward, distance-based approach. It partitions observations into K clusters by iteratively assigning them to the nearest centroids. This section details the K-Means algorithm’s fundamentals, plus considerations like standardization and the Elbow Method for choosing K.

K-MEANS FUNDAMENTALS

Learning Objectives

• Cluster data into K groups using distance-based assignment

• Summarize and profile clusters for actionable insights

Indicative Content

• Algorithm Steps

  • Initialize centroids → assign points → recalc centroids → converge

• Scaling

  • Importance of standardizing variables to prevent magnitude bias

K-MEANS IMPLEMENTATION & ELBOW METHOD

Learning Objectives

• Use the Elbow Method to find an appropriate K

Indicative Content

• Elbow Plot

  • WCSS (within-cluster sum of squares)/inertia vs. K

• Interpretation

  • Observing the “bend” to choose optimal clusters

TOOLS & METHODOLOGIES (NON-HIERARCHICAL CLUSTERING)

• Python

  • sklearn.cluster.KMeans for centroid-based segmentation

• Workflow

  • Scale data → set initial K → run K-Means → evaluate WCSS

• Evaluation

  • Elbow Plot to refine K

  • Cluster profiles for meaningful segmentation