Chapter 12

Title: Grouping Higher Education Students with RapidMiner

Summary

Chapter 12 applies clustering to automatically group higher education students. The dataset corresponds to the one already described in Chapter 10, but now the task is to find groups of similarly performing students, which is achieved with automated clustering techniques. The attributes describing the students may have missing values and different scales. Hence data pre-processing steps are used to replace missing values and to normalize the attribute values to identical value ranges. A parameter loop automatically selects and evaluates the performance of several clustering techniques including k-Means, k-Medoids, Support Vector Clustering (SVC), and DBSCAN.

Table of Contents

12.0 Overview
12.1 Introduction
12.2 Related Work
12.3 Using RapidMiner for Clustering Higher Education Students
12.3.1 Data
12.3.2 Process for Automatic Evaluation of Clustering Algorithms
12.3.3 Results and Discussion
12.4 Conclusion
12.4 Bibliography

Dataset & Processes: Click here to download