Cluster Analysis

In this course, you will learn the basics of cluster analysis, one of the most popular data mining methods for the discovery of patterns in learning data, and its application in learning analytics.

Cluster analysis enables the identification of common, archetypal patterns of student interactions, which can lead to better understanding of student learning behaviors and provision of personalized feedback and interventions.

This course will have a strong hands-on component, as you will learn how to conduct a cluster analysis using the popular Weka data mining toolkit.

We will cover K-means and Hierarchical clustering techniques, which are two simple, yet widely used, cluster analysis methods. We will also review some of the published learning analytics studies that adopted cluster analysis and learn how to interpret the cluster analysis results.

Finally, we will also examine some of the more advanced techniques and identify certain practical challenges with cluster analysis, such as the selection of the optimal number of clusters and the validation of cluster analysis results.

Week 1: Introduction
Lectures:

1. Introduction to unsupervised machine learning methods
2. Introduction to clustering
3. Overview of clustering uses for learning analytics

Labs:

1. Introduction to Weka toolkit

Week 2: Overview of k-means and hierarchical clustering methods
Lectures:

1. K-means clustering theory
2. K-means full example
3. Hierarchical clustering theory
4. Hierarchical clustering full example

Labs:

1. Conducting k-means clustering using Weka
2. Conducting hierarchical clustering using Weka

Week 3: Practical considerations
Lectures:

1. How to choose the number of clusters
2. How to interpret clustering results
3. Overview of more advanced clustering methods

Labs:

1. Real-world cluster analysis walkthrough