Block GTM: Incorporating prior knowledge of covariance structure in data visualisation

Martin Schroeder, Ian T. Nabney, Dan Cornford

Research output: Working paperTechnical report

Abstract

Visualising data for exploratory analysis is a big challenge in scientific and engineering domains where there is a need to gain insight into the structure and distribution of the data. Typically, visualisation methods like principal component analysis and multi-dimensional scaling are used, but it is difficult to incorporate prior knowledge about structure of the data into the analysis. In this technical report we discuss a complementary approach based on an extension of a well known non-linear probabilistic model, the Generative Topographic Mapping. We show that by including prior information of the covariance structure into the model, we are able to improve both the data visualisation and the model fit.
Original languageEnglish
Place of PublicationBirmingham
PublisherAston University
Number of pages29
ISBN (Print)NCRG/2008/006
Publication statusPublished - 25 Sep 2008

Fingerprint

Data visualization
Principal component analysis
Visualization
Statistical Models

Keywords

  • Including prior information of the covariance structure
  • Generative Topographic Mapping
  • Improving Data Visualisation

Cite this

Schroeder, M., Nabney, I. T., & Cornford, D. (2008). Block GTM: Incorporating prior knowledge of covariance structure in data visualisation. Birmingham: Aston University.
Schroeder, Martin ; Nabney, Ian T. ; Cornford, Dan. / Block GTM: Incorporating prior knowledge of covariance structure in data visualisation. Birmingham : Aston University, 2008.
@techreport{035c0e7e5974408d96171464e93ccd44,
title = "Block GTM: Incorporating prior knowledge of covariance structure in data visualisation",
abstract = "Visualising data for exploratory analysis is a big challenge in scientific and engineering domains where there is a need to gain insight into the structure and distribution of the data. Typically, visualisation methods like principal component analysis and multi-dimensional scaling are used, but it is difficult to incorporate prior knowledge about structure of the data into the analysis. In this technical report we discuss a complementary approach based on an extension of a well known non-linear probabilistic model, the Generative Topographic Mapping. We show that by including prior information of the covariance structure into the model, we are able to improve both the data visualisation and the model fit.",
keywords = "Including prior information of the covariance structure, Generative Topographic Mapping, Improving Data Visualisation",
author = "Martin Schroeder and Nabney, {Ian T.} and Dan Cornford",
year = "2008",
month = "9",
day = "25",
language = "English",
isbn = "NCRG/2008/006",
publisher = "Aston University",
type = "WorkingPaper",
institution = "Aston University",

}

Schroeder, M, Nabney, IT & Cornford, D 2008 'Block GTM: Incorporating prior knowledge of covariance structure in data visualisation' Aston University, Birmingham.

Block GTM: Incorporating prior knowledge of covariance structure in data visualisation. / Schroeder, Martin; Nabney, Ian T.; Cornford, Dan.

Birmingham : Aston University, 2008.

Research output: Working paperTechnical report

TY - UNPB

T1 - Block GTM: Incorporating prior knowledge of covariance structure in data visualisation

AU - Schroeder, Martin

AU - Nabney, Ian T.

AU - Cornford, Dan

PY - 2008/9/25

Y1 - 2008/9/25

N2 - Visualising data for exploratory analysis is a big challenge in scientific and engineering domains where there is a need to gain insight into the structure and distribution of the data. Typically, visualisation methods like principal component analysis and multi-dimensional scaling are used, but it is difficult to incorporate prior knowledge about structure of the data into the analysis. In this technical report we discuss a complementary approach based on an extension of a well known non-linear probabilistic model, the Generative Topographic Mapping. We show that by including prior information of the covariance structure into the model, we are able to improve both the data visualisation and the model fit.

AB - Visualising data for exploratory analysis is a big challenge in scientific and engineering domains where there is a need to gain insight into the structure and distribution of the data. Typically, visualisation methods like principal component analysis and multi-dimensional scaling are used, but it is difficult to incorporate prior knowledge about structure of the data into the analysis. In this technical report we discuss a complementary approach based on an extension of a well known non-linear probabilistic model, the Generative Topographic Mapping. We show that by including prior information of the covariance structure into the model, we are able to improve both the data visualisation and the model fit.

KW - Including prior information of the covariance structure

KW - Generative Topographic Mapping

KW - Improving Data Visualisation

M3 - Technical report

SN - NCRG/2008/006

BT - Block GTM: Incorporating prior knowledge of covariance structure in data visualisation

PB - Aston University

CY - Birmingham

ER -

Schroeder M, Nabney IT, Cornford D. Block GTM: Incorporating prior knowledge of covariance structure in data visualisation. Birmingham: Aston University. 2008 Sep 25.