Dynamic Latent Classification Model: Towards a More Expressive Model for Dynamic Classification

Shengtong Zhong, Ana M. Martínez, Thomas Dyhre Nielsen, Helge Langseth

Research output: Contribution to conference without publisher/journalPaper without publisher/journalResearchpeer-review

Abstract

Monitoring a complex process often involves keeping an eye on hundreds or thousands of sensors to determine whether or not the process is under control. We have been working with dynamic data from an oil production facility in the North sea, where unstable situations should be identified as soon as possible. Motivated by this problem setting, we propose a generative model for dynamic classification in continuous domains. At each time point the model can be seen as combining a naive Bayes model with a mixture of factor analyzers (FA). The latent variables of the FA are used to capture the dynamics in the process as well as modeling dependences between attributes.
Original languageEnglish
Publication date2010
Publication statusPublished - 2010
EventThe Florida Artificial Intelligence Research Society Conference - Daytona Beach, Florida, United States
Duration: 19 May 2010 → …
Conference number: 23

Conference

ConferenceThe Florida Artificial Intelligence Research Society Conference
Number23
CountryUnited States
CityDaytona Beach, Florida
Period19/05/2010 → …

Fingerprint

Monitoring
Sensors
Oils

Cite this

Zhong, S., Martínez, A. M., Nielsen, T. D., & Langseth, H. (2010). Dynamic Latent Classification Model: Towards a More Expressive Model for Dynamic Classification. Paper presented at The Florida Artificial Intelligence Research Society Conference, Daytona Beach, Florida, United States.
Zhong, Shengtong ; Martínez, Ana M. ; Nielsen, Thomas Dyhre ; Langseth, Helge. / Dynamic Latent Classification Model : Towards a More Expressive Model for Dynamic Classification. Paper presented at The Florida Artificial Intelligence Research Society Conference, Daytona Beach, Florida, United States.
@conference{238809405c6e11df8573000ea68e967b,
title = "Dynamic Latent Classification Model: Towards a More Expressive Model for Dynamic Classification",
abstract = "Monitoring a complex process often involves keeping an eye on hundreds or thousands of sensors to determine whether or not the process is under control. We have been working with dynamic data from an oil production facility in the North sea, where unstable situations should be identified as soon as possible. Motivated by this problem setting, we propose a generative model for dynamic classification in continuous domains. At each time point the model can be seen as combining a naive Bayes model with a mixture of factor analyzers (FA). The latent variables of the FA are used to capture the dynamics in the process as well as modeling dependences between attributes.",
author = "Shengtong Zhong and Mart{\'i}nez, {Ana M.} and Nielsen, {Thomas Dyhre} and Helge Langseth",
year = "2010",
language = "English",
note = "null ; Conference date: 19-05-2010",

}

Zhong, S, Martínez, AM, Nielsen, TD & Langseth, H 2010, 'Dynamic Latent Classification Model: Towards a More Expressive Model for Dynamic Classification' Paper presented at, Daytona Beach, Florida, United States, 19/05/2010, .

Dynamic Latent Classification Model : Towards a More Expressive Model for Dynamic Classification. / Zhong, Shengtong; Martínez, Ana M.; Nielsen, Thomas Dyhre; Langseth, Helge.

2010. Paper presented at The Florida Artificial Intelligence Research Society Conference, Daytona Beach, Florida, United States.

Research output: Contribution to conference without publisher/journalPaper without publisher/journalResearchpeer-review

TY - CONF

T1 - Dynamic Latent Classification Model

T2 - Towards a More Expressive Model for Dynamic Classification

AU - Zhong, Shengtong

AU - Martínez, Ana M.

AU - Nielsen, Thomas Dyhre

AU - Langseth, Helge

PY - 2010

Y1 - 2010

N2 - Monitoring a complex process often involves keeping an eye on hundreds or thousands of sensors to determine whether or not the process is under control. We have been working with dynamic data from an oil production facility in the North sea, where unstable situations should be identified as soon as possible. Motivated by this problem setting, we propose a generative model for dynamic classification in continuous domains. At each time point the model can be seen as combining a naive Bayes model with a mixture of factor analyzers (FA). The latent variables of the FA are used to capture the dynamics in the process as well as modeling dependences between attributes.

AB - Monitoring a complex process often involves keeping an eye on hundreds or thousands of sensors to determine whether or not the process is under control. We have been working with dynamic data from an oil production facility in the North sea, where unstable situations should be identified as soon as possible. Motivated by this problem setting, we propose a generative model for dynamic classification in continuous domains. At each time point the model can be seen as combining a naive Bayes model with a mixture of factor analyzers (FA). The latent variables of the FA are used to capture the dynamics in the process as well as modeling dependences between attributes.

M3 - Paper without publisher/journal

ER -

Zhong S, Martínez AM, Nielsen TD, Langseth H. Dynamic Latent Classification Model: Towards a More Expressive Model for Dynamic Classification. 2010. Paper presented at The Florida Artificial Intelligence Research Society Conference, Daytona Beach, Florida, United States.