Machine Learning (ML) · Intermediate level

Intermediate level · Module 4 of 6 · 15 guided hours

Unsupervised Learning

Find groups and structure in unlabelled data with clustering and dimensionality reduction, and turn them into useful insights.

Learning objectives

  • Segment data with k-means and choose the number of clusters
  • Use hierarchical clustering and DBSCAN where they fit better
  • Reduce dimensions with PCA for visualisation and modelling

Topics

  • k-means
  • elbow and silhouette methods
  • hierarchical clustering
  • DBSCAN and outliers
  • PCA
  • interpreting and validating clusters

Lessons

  1. Lesson 1 k-Means Clustering

    Group similar records with k-means and choose a sensible number of clusters.

    k-meanscentroidinertia and elbowsilhouette score

  2. Lesson 2 Other Clustering Methods

    Use hierarchical clustering and DBSCAN when k-means' assumptions do not fit, and detect outliers.

    hierarchical clusteringdendrogramDBSCANnoise points

  3. Lesson 3 Dimensionality Reduction with PCA

    Summarise many correlated features with principal component analysis for visualisation and modelling.

    principal componentsexplained variance ratio2-D projectioninformation loss

Practical activity

Segment mobile-money agents by transaction behaviour, name and describe each segment and propose a different support action for each.

Expected outcomes

  • Meaningful, interpretable segments
  • Appropriate use of PCA

Assessment

Segmentation report marked on method, cluster validation and business interpretation.

Taught live by an instructor — online or in person. Free course lessons are for self-study and do not lead to a certificate — see registration options.