Experimental analysis of the q-matrix method in knowledge discovery

Authors:
Tiffany Barnes;Donald Bitzer;Mladen Vouk
Affiliations:
Computer Science Dept., University of North Carolina at Charlotte, Charlotte, NC;Computer Science Dept., North Carolina State University, Raleigh, NC;Computer Science Dept., North Carolina State University, Raleigh, NC
Venue:
ISMIS'05 Proceedings of the 15th international conference on Foundations of Intelligent Systems
Year:
2005

Citing 2
Cited 0

Data Mining by Means of Binary Representation: A Model for Similarity and Clustering

Information Systems Frontiers
Student Modeling from Conversational Test Data: A Bayesian Approach Without Priors

ITS '98 Proceedings of the 4th International Conference on Intelligent Tutoring Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

The q-matrix method, a new method for data mining and knowledge discovery, is compared with factor analysis and cluster analysis in analyzing fourteen experimental data sets. This method creates a matrix-based model that extracts latent relationships among observed binary variables. Results show that the q-matrix method offers several advantages over factor analysis and cluster analysis for knowledge discovery. The q-matrix method can perform fully unsupervised clustering, where the number of clusters is not known in advance. It also yields better error rates than factor analysis, and is comparable in error to cluster analysis. The q-matrix method also allows for automatic interpretation of the data sets. These results suggest that the q-matrix method can be an important tool in automated knowledge discovery.