A new framework for clustering algorithm evaluation in the domain of functional genomics

Authors:
AnYuan Guo
Affiliations:
University of Massachusetts Amherst, Amherst, MA
Venue:
Proceedings of the 2004 ACM symposium on Applied computing
Year:
2004

Citing 5
Cited 0

Fusion, propagation, and structuring in belief networks

Artificial Intelligence
Probabilistic inference and influence diagrams

Operations Research
A Bayesian Method for the Induction of Probabilistic Networks from Data

Machine Learning
A tutorial on learning with Bayesian networks

Learning in graphical models
Using Bayesian networks to analyze expression data

RECOMB '00 Proceedings of the fourth annual international conference on Computational molecular biology

Quantified Score

Hi-index	0.00

Visualization

Abstract

Clustering algorithms are widely used in the computational analysis of microarray data. However, due to the lack of domain knowledge, it is often difficult to judge their performance. In this paper, we introduce a new framework for the evaluation of clustering algorithms in application to regulatory pathway reconstruction. A pilot study was conducted on the hierarchical clustering algorithm for which we obtained qualitative characterizations of the number of samples needed as well as the denseness of the subnetwork required to achieve accurate partition. For experimental scientists, this evaluation framework provides a method to select and calibrate clustering algorithms. It can also provide a confidence measure to the results of a clustering algorithm when certain restrictions on the experimental setup, such as the number of samples available, are known in advance.